| Title: | Diagnostic Analysis Using Forward Search Procedure for Various Models |
|---|---|
| Description: | Identifies potential data outliers and their impact on estimates and analyses. Uses the forward search approach of Atkinson and Riani, "Robust Diagnostic Regression Analysis", 2000,<ISBN: o-387-95017-6> to prepare descriptive statistics of a dataset that is to be analyzed by stats::lm(), stats::glm(), stats::nls(), nlme::lme() or survival::coxph(). Includes graphics functions to display the descriptive statistics. |
| Authors: | William Fairweather [aut, cre] |
| Maintainer: | William Fairweather <[email protected]> |
| License: | GPL (>= 3) |
| Version: | 6.2.0 |
| Built: | 2024-09-12 10:16:04 UTC |
| Source: | CRAN |
Identifies potential data outliers and their impact on estimates and analyses. Uses the forward search approach of Atkinson and Riani, "Robust Diagnostic Regression Analysis", 2000,<ISBN: o-387-95017-6> to prepare descriptive statistics of a dataset that is to be analyzed by stats::lm(), stats::glm(), stats::nls(), nlme::lme() or survival::coxph(). Includes graphics functions to display the descriptive statistics.
The DESCRIPTION file:
| Package: | forsearch |
| Title: | Diagnostic Analysis Using Forward Search Procedure for Various Models |
| Version: | 6.2.0 |
| Authors@R: | person("William","Fairweather", email = "[email protected]", role = c("aut", "cre")) |
| Description: | Identifies potential data outliers and their impact on estimates and analyses. Uses the forward search approach of Atkinson and Riani, "Robust Diagnostic Regression Analysis", 2000,<ISBN: o-387-95017-6> to prepare descriptive statistics of a dataset that is to be analyzed by stats::lm(), stats::glm(), stats::nls(), nlme::lme() or survival::coxph(). Includes graphics functions to display the descriptive statistics. |
| Depends: | R (>= 4.2) |
| License: | GPL (>= 3) |
| SystemRequirements: | gmp (>= 4.1) |
| Encoding: | UTF-8 |
| RoxygenNote: | 7.2.3 |
| Imports: | Hmisc(>= 4.7-0), Cairo(>= 1.6-0), formula.tools(>= 1.7.0), ggplot2(>= 3.4.0), nlme(>= 3.1-157), survival(>= 3.4), tibble(>= 3.1.8) |
| Suggests: | rmarkdown, knitr |
| VignetteBuilder: | knitr |
| NeedsCompilation: | no |
| Packaged: | 2024-07-13 14:44:04 UTC; NO |
| Author: | William Fairweather [aut, cre] |
| Maintainer: | William Fairweather <[email protected]> |
| Repository: | CRAN |
| Date/Publication: | 2024-07-13 15:00:02 UTC |
Index of help topics:
aStep1 Create Set of Observation Numbers in Step 1 for
Linear Model Analysis
aStep2 Update Observation Set in Step 2
bStep1 Create Set of Observation Numbers in Step 1 for
Linear Mixed Effects Model Analysis
bStep2 Update Observation Numbers in Step 2
cStep1 Create Set of Observation Numbers in Step 1 for
Cox Proportional Hazards Model Analysis
cStep2 Update Observation Set in Step 2
forsearch-package Diagnostic Analysis Using Forward Search
Procedure for Various Models Diagnostic
Analysis Using Forward Search Procedure for
Various Models
forsearch_cph Create Statistics Of Forward Search in a Cox
Proportional Hazard Database
forsearch_glm Create Statistics of Forward Search in a
Generalized Linear Model Database
forsearch_lm Create Statistics Of Forward Search in a Linear
Model Database
forsearch_lme Create Statistics Of Forward Search For a
Linear Mixed Effects Database
forsearch_nls Create Statistics Of Forward Search in a
Nonlinear Model Database
identifyCoeffs Index To Identify Fixed and Random Coefficients
To Appear Together on Plot
identifyFixedCoeffs Index To Identify Fixed Coefficients To Appear
Together on Plot
picksome Structured Sets of Random Samples of
Observations
plotdiag.AICX Plot Diagnostic AIC Statistics
plotdiag.ANOX2 Plot Diagnostic Statistics of Analysis of
Variance Tables
plotdiag.Cook Plot Diagnostic Statistics of Modified Cook's
Distance
plotdiag.Wald Plot Diagnostic Statistics of Wald Test Output
of COXPH Function
plotdiag.allgraphs Execute All Plotting Functions For a Select
Forsearch Object
plotdiag.blind.fixed Plot Diagnostic Statistics of Fixed
Coefficients for Blinded Dataset
plotdiag.deviance.residuals
Plot Diagnostic Statistics Of Deviance
Residuals
plotdiag.deviances Plot Diagnostic Deviance Statistics
plotdiag.fit3 Plot Diagnostic Statistics of AIC, BIC, and Log
Likelihood
plotdiag.leverage Plot Diagnostic Statistics Of Leverage
plotdiag.loglik Plot Diagnostic Statistics of LOGLIK Output of
COXPH Function
plotdiag.lrt Plot Diagnostic Statistics of Likelihood Ratio
Test of COXPH Function
plotdiag.params.fixed Plot Diagnostic Statistics of Fixed
Coefficients
plotdiag.params.random
Plot Diagnostic Statistics Of Random
Coefficients
plotdiag.phihatx Plot Diagnostic PhiHat Statistics
plotdiag.residuals Plot Diagnostic Statistics Of Residuals Or
Squared Residuals
plotdiag.s2 Plot Diagnostic Statistics Of Residual
Variation
plotdiag.tstats Plot Diagnostic T Statistics
search.history Create Tabular History Of Forward Search
showme Display Abbreviated Output of FORSEARCH_xxx
Function
variablelist Identify Level(s) to Which Each Factor
Observation Belongs
Further information is available in the following vignettes:
Exploring-the-Search-History |
Exploring the Search History (source, pdf) |
How-many-observations-are-needed-and-where-do-we-get-them |
How Many Observations are Needed and Where Do We Get Them? (source, pdf) |
Quality-control-of-the-dataset-using-the-forward-search |
Quality control of the dataset using the forward search (source, pdf) |
Ensure that data frame has a leading column of observation numbers. Run forsearch_xxx to create a file of diagnostic statistics to be used as input to such plotting functions as plotdiag.residuals, plotdiag.params.fixed, plotdiag.params.random, plotdiag.s2, plotdiag,leverage, and plotdiag.Cook. The file of diagnostic statistics can be voluminous, and the utility function showme displays the output more succinctly. Plotting of statistics for fixed and for random coefficients is limited by graphical restraints in some cases. The function identifyCoeffs provides a set of indexing codes so that plotdiag.params.random can display diagnostics for selected fixed or random model parameters. The function identifyFixedCoeffs does the same for lm models.
William R. Fairweather, Flower Valley Consulting, Inc., Silver Spring MD USA William Fairweather [aut, cre]
Maintainer: William Fairweather <[email protected]> William R. Fairweather <wrf343 at flowervalleyconsulting.com>
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000. Pinheiro, JC and DM Bates. Mixed-Effects Models in S and S-Plus, Springer, New York, 2000.
Derives the first set of observation numbers for forsearch in linear models
aStep1(yesfactor, df1, df1.ls, inner.rank, initial.sample, formula, ycol, nopl, b.d)aStep1(yesfactor, df1, df1.ls, inner.rank, initial.sample, formula, ycol, nopl, b.d)
yesfactor |
Logical. TRUE if there are factors in the X matrix |
df1 |
Data frame being analyzed by forward search. |
df1.ls |
List, each element of which is a factor subset of df1 |
inner.rank |
Rank of X matrix of lm analysis on entire database |
initial.sample |
Number of random samples from which to take set of initial observations |
formula |
Fixed parameter formula of lm function |
ycol |
Response column number |
nopl |
Number of observations per level of combined factor variables |
b.d |
Index of point to begin diagnostic listings |
Support function, usually not called independently
Produces set of observation numbers for Step 1. Accounts for presence of factors in the dataset
Presence of Observation column has no effect on outcome
William R. Fairweather
Derives the set of observation numbers for forsearch in Step 2 for linear models
aStep2(yesfactor, form.A2, finalm, rimbs, dfa2, ycol, mstart, rnk, b.d)aStep2(yesfactor, form.A2, finalm, rimbs, dfa2, ycol, mstart, rnk, b.d)
yesfactor |
True or False for presence of factors |
form.A2 |
Formula for analysis of entire dataset |
finalm |
See VALUE above. finalm argument is the same but only for Step 1 values |
rimbs |
List, each element is a matrix of obs numbers and corresponding subset codes |
dfa2 |
Data frame being analyzed by forward search. Presence of Observation column has no effect on output |
ycol |
Response column number, including 1 for Observation |
mstart |
Number of first subset to be defined in Step 2 |
rnk |
Rank of X matrix. For factors, this is rank with factors removed. |
b.d |
Number at which to begin diagnostic listings |
Support function, usually not called independently
Vector of integers corresponding to observation numbers
William R. Fairweather
Derives the first set of observation numbers for forsearch in linear mixed effects models
bStep1(yesfactor, df1, df1.ls, inner.rank, initial.sample, formula, randform, ycol, nopl, b.d)bStep1(yesfactor, df1, df1.ls, inner.rank, initial.sample, formula, randform, ycol, nopl, b.d)
yesfactor |
Logical. TRUE if there are factors in the X matrix |
df1 |
Data frame being analyzed by forward search. |
df1.ls |
List, each element of which is a factor subset of df1 |
inner.rank |
Rank of X matrix of lme analysis on entire database |
initial.sample |
Number of random samples from which to take set of initial observations |
formula |
Two-sided fixed parameter formula of lme function |
randform |
One-sided random effects formula |
ycol |
Response column number |
nopl |
Number of observations per level of combined factor variables |
b.d |
Index of point to begin diagnostic listings |
Support function, usually not called independently
Produces set of observation numbers for Step 1. Accounts for presence of factors in the dataset
Presence of Observation column has no effect on outcome
William R. Fairweather
Derives the set of Step 2 observation numbers for forsearch in linear mixed effects models
bStep2(f2, dfa2, randm2, ms, finalm, fbg, b.d, rnk2, ycol)bStep2(f2, dfa2, randm2, ms, finalm, fbg, b.d, rnk2, ycol)
f2 |
Fixed parameter formula |
dfa2 |
Complete data set with factor subset identification codes |
randm2 |
Random parameter formula |
ms |
Number of observations beginning Step 2 |
finalm |
List of expanding subset observation numbers |
fbg |
List of observation numbers by factor subgroup |
b.d |
Indicator of place in code to begin diagnostic printouts |
rnk2 |
Rank of linear regression with factor variables eliminated |
ycol |
Column number of response variable |
Support function, usually not called independently
List of expanding number sets corresponding to observation numbers
William R. Fairweather
Derives the first set of observation numbers for forsearch in Cox Proportional Hazards models
cStep1(yesfactor, df1, df1.ls, inner.rank, initial.sample, formula, f.e, ycol, nopl, b.d)cStep1(yesfactor, df1, df1.ls, inner.rank, initial.sample, formula, f.e, ycol, nopl, b.d)
yesfactor |
Logical. TRUE if there are factors in the X matrix |
df1 |
Data frame being analyzed by forward search. |
df1.ls |
List, each element of which is a factor subset of df1 |
inner.rank |
Rank of X matrix of lm analysis on entire database |
initial.sample |
Number of random samples from which to take set of initial observations |
formula |
Fixed parameter formula of lm function |
f.e |
Right-hand side of formula for Surv function |
ycol |
Response column number |
nopl |
Number of observations per level of combined factor variables |
b.d |
Index of point to begin diagnostic listings |
Support function, usually not called independently
Produces set of observation numbers for Step 1. Accounts for presence of factors in the dataset
Presence of Observation column has no effect on outcome
William R. Fairweather
Derives the set of observation numbers for step 2 for forsearch in Cox proportional hazard models
cStep2(f.e, finalm, dfa2, ms, rnk2, ss, b.d)cStep2(f.e, finalm, dfa2, ms, rnk2, ss, b.d)
f.e |
Right hand side of formula |
finalm |
List of rows in model at each stage |
dfa2 |
Complete data frame with factor subset indicator codes |
ms |
Number of observations in first stage of Step 2 |
rnk2 |
Rank of linear analysis with factor variables removed |
ss |
NULL or vector of observation numbers manually entered into Step 1 |
b.d |
Indicator of starting point for diagnostic listings |
Support function, usually not called independently
Vector of expanding number sets corresponding to observation numbers
William R. Fairweather
Prepares summary statistics at each stage of forward search for subsequent plotting.
forsearch_cph(alldata, formula.rhs, initial.sample=1000, n.obs.per.level=1, skip.step1=NULL, ties = "efron", maxdisturb=0.01, proportion=TRUE, unblinded=TRUE, begin.diagnose= 100, verbose=TRUE)forsearch_cph(alldata, formula.rhs, initial.sample=1000, n.obs.per.level=1, skip.step1=NULL, ties = "efron", maxdisturb=0.01, proportion=TRUE, unblinded=TRUE, begin.diagnose= 100, verbose=TRUE)
alldata |
Data frame containing variables 'Observation', 'event.time', 'status', and independent variables, in that order |
formula.rhs |
Character vector of names of independent variables in model |
initial.sample |
Number of observations in Step 1 of forward search |
n.obs.per.level |
Number of observations per level of (possibly crossed) factor levels to include in Step 1 |
skip.step1 |
NULL or a vector of integers for observations to be included in Step 1 |
ties |
Method for handling ties in event time; = "efron", "breslow", or "exact"; see survival::coxph |
maxdisturb |
Maximum amount to add randomly to event.time to prevent ties. |
proportion |
TRUE causes evaluation of proportionality of Cox regression |
unblinded |
TRUE causes printing of presumed analysis structure |
begin.diagnose |
Numeric. Indicates where in code to begin printing diagnostics. 0 prints all; 100 prints none |
verbose |
TRUE causes function identifier display before and after run |
LIST
Rows in stage |
Observation numbers of rows included at each stage |
Number of model parameters |
Number of fixed coefficients in Cox model |
Fixed parameter estimates |
Vector of parameter estimates at each stage |
Wald Test |
Vector of Wald tests at each stage |
Proportionality Test |
Result of Cox proportionality test, if run |
LogLikelihood |
Vector of null and overall coefficients log likelihoods at each stage |
Likelihood ratio test |
Vector of LRTs at each stage |
Leverage |
Matrix of leverage of each observation at each stage |
Call |
Call to this function |
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
## Not run: {# Forsearch for Cox Proportional Survival, including Step 1 veteran <- survival::veteran veteran <- veteran[order(veteran$celltype),] veteranx <- veteran[,c(3,4,1,2)] veteranx$trt <- as.factor(veteranx$trt) dimv <- dim(veteran)[1] Observation <- 1:dimv veteranx <- data.frame(Observation,veteranx) names(veteranx)[2] <- "event.time" form.1 <- "trt + celltype" forskip <- NULL # forskip <- c(12, 23, 38, 71, 91, 104, 116, 130, 31, 73, 62, 76) cphtest1a.out <- forsearch_cph(alldata=veteranx, formula.rhs=form.1, n.obs.per.level=2, skip.step1=forskip, ties="efron", unblinded=TRUE, initial.sample=467, begin.diagnose = 100, verbose = TRUE) } {# Same, but skipping Step 1. forskip <- c(12, 6, 31, 23, 38, 62, 71, 73, 91, 84, 104, 101, 116, 125,128,76) cphtest1b.out <- forsearch_cph(alldata=veteranx, formula.rhs=form.1, n.obs.per.level=2, skip.step1=forskip, ties="efron", unblinded=TRUE, initial.sample=467, begin.diagnose = 100, verbose = TRUE) } ## End(Not run)## Not run: {# Forsearch for Cox Proportional Survival, including Step 1 veteran <- survival::veteran veteran <- veteran[order(veteran$celltype),] veteranx <- veteran[,c(3,4,1,2)] veteranx$trt <- as.factor(veteranx$trt) dimv <- dim(veteran)[1] Observation <- 1:dimv veteranx <- data.frame(Observation,veteranx) names(veteranx)[2] <- "event.time" form.1 <- "trt + celltype" forskip <- NULL # forskip <- c(12, 23, 38, 71, 91, 104, 116, 130, 31, 73, 62, 76) cphtest1a.out <- forsearch_cph(alldata=veteranx, formula.rhs=form.1, n.obs.per.level=2, skip.step1=forskip, ties="efron", unblinded=TRUE, initial.sample=467, begin.diagnose = 100, verbose = TRUE) } {# Same, but skipping Step 1. forskip <- c(12, 6, 31, 23, 38, 62, 71, 73, 91, 84, 104, 101, 116, 125,128,76) cphtest1b.out <- forsearch_cph(alldata=veteranx, formula.rhs=form.1, n.obs.per.level=2, skip.step1=forskip, ties="efron", unblinded=TRUE, initial.sample=467, begin.diagnose = 100, verbose = TRUE) } ## End(Not run)
Prepares summary statistics at each stage of forward search for subsequent plotting. Forward search is conducted in three steps: Step 1 to identify minimal set of observations to estimate unknown parameters, and Step 2 to add one observation at each stage such that observations in the set are best fitting at that stage. A preliminary step (Step 0) contains code for pre-processing of the data.
forsearch_glm(initial.sample=1000, response.cols, indep.cols, family, formula=NULL, binomialrhs=NULL, formula.cont.rhs, data, n.obs.per.level = 1, estimate.phi = TRUE, skip.step1=NULL, unblinded=TRUE, begin.diagnose=100, verbose=TRUE)forsearch_glm(initial.sample=1000, response.cols, indep.cols, family, formula=NULL, binomialrhs=NULL, formula.cont.rhs, data, n.obs.per.level = 1, estimate.phi = TRUE, skip.step1=NULL, unblinded=TRUE, begin.diagnose=100, verbose=TRUE)
initial.sample |
Number of random sets of observations in Step 1 of forward search |
response.cols |
Vector of column numbers (1 or 2) of responses and nonresponses (if binomial) |
indep.cols |
Column number(s) of independent variables |
family |
Error distribution and link |
formula |
Formula relating response to independent variables. Required except for family=binomial |
binomialrhs |
Quoted character.Right-hand side of formula. Required for family=binomial |
formula.cont.rhs |
Quoted character.Right-hand side of formula, omitting factor variables. Required for all families |
data |
Name of database |
n.obs.per.level |
Number of observations per level of (possibly crossed) factor levels |
estimate.phi |
TRUE causes phi to be estimated; FALSE causes phi to be set = 1 |
skip.step1 |
NULL, or vector of observation numbers to include at end of Step 1 |
unblinded |
TRUE allows print of formula of analysis function |
begin.diagnose |
Numeric. Indicates where in code to begin printing diagnostics. 0 prints all; 100 prints none |
verbose |
TRUE causes function identifier to display before and after run |
Step 2 is determined by the results of Step 1, which itself is random. So, it is possible to reproduce the entire run by using the skip.step1 argument. Inner subgroups are produced by presence of categorical variables. Current version assumes independent variables are all continuous.
LIST
Rows in stage |
Observation numbers of rows included at each stage |
Family |
Family and link |
Number of model parameters |
Number of fixed effect parameters |
Fixed parameter estimates |
Matrix of parameter estimates at each stage |
Residual deviance |
Vector of deviances |
Null deviance |
Vector of null deviances |
PhiHat |
Vector of values of phi parameter |
Deviance residuals and augments |
Deviance residuals with indication of whether each is included in fit |
AIC |
Vector of AIC values |
Leverage |
Matrix of leverage of each observation at each stage |
Call |
Call to this function |
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
# Train deaths (Atkinson and Riani, 2000) with Rolling Stock as a factor Observation<-1:67 Month<-c(9,8,3,1,10,6,7,1,8,4,3,3,12,11,10,9,9,4,3,12,12,10,7,2,12,2,12,12,12, 11,3,10,4,2,12,12,9,11,1,10,8,6,1,10,6,12,8,4,9,6,12,10,7,2,5,12,5,5,4,3,1, 9,11,9,7,3,2) Year<-c(97,96,96,95,94,94,91,91,90,89,89,89,88,88,87,86,86,86,86,84,84,84,84,84, 83,83,82,81,81,80,80,79,79,79,78,78,77,76,76,75,75,75,75,74,74,73,73,73,72, 72,71,71,71,71,70,69,69,69,69,69,69,68,67,67,67,67,67) RollingStock<-c(2,2,3,2,1,1,1,1,2,3,1,1,1,2,1,2,1,3,2,2,1,2,2,3,1,2,1,1,2,3,1, 1,1,1,1,1,1,3,3,2,3,1,2,3,1,1,1,3,3,1,3,3,1,1,1,2,1,1,2,1,1,1,1,1,1,1,1) RollingStock <- as.factor(RollingStock) Traffic<-c(0.436,0.424,0.424,0.426,0.419,0.419,0.439,0.439,0.431,0.436,0.436, 0.436,0.443,0.443,0.397,0.414,0.414,0.414,0.414,0.389,0.389,0.389,0.389, 0.389,0.401,0.401,0.372,0.417,0.417,0.43,0.43,0.426,0.426,0.426,0.43,0.43, 0.425,0.426,0.426,0.436,0.436,0.436,0.436,0.452,0.452,0.433,0.433,0.433, 0.431,0.431,0.444,0.444,0.444,0.444,0.452,0.447,0.447,0.447,0.447,0.447, 0.447,0.449,0.459,0.459,0.459,0.459,0.459) Deaths<-c(7,1,1,1,5,2,4,2,1,1,2,5,35,1,4,1,2,1,1,3,1,3,13,2,1,1,1,4,1,2,1,5,7, 1,1,3,2,1,2,1,2,6,1,1,1,10,5,1,1,6,3,1,2,1,2,1,1,6,2,2,4,2,49,1,7,5,9) train2022 <- data.frame(Observation, Year, RollingStock, Traffic, Deaths) forsearch_glm(initial.sample = 100, response.cols = 5, indep.cols = 2:4, formula=Deaths~Year + RollingStock + Traffic, formula.cont.rhs="Year + Traffic", family = poisson("log"), data = train2022, n.obs.per.level = 1, estimate.phi = TRUE, skip.step1 = NULL, unblinded = TRUE, begin.diagnose=100)# Train deaths (Atkinson and Riani, 2000) with Rolling Stock as a factor Observation<-1:67 Month<-c(9,8,3,1,10,6,7,1,8,4,3,3,12,11,10,9,9,4,3,12,12,10,7,2,12,2,12,12,12, 11,3,10,4,2,12,12,9,11,1,10,8,6,1,10,6,12,8,4,9,6,12,10,7,2,5,12,5,5,4,3,1, 9,11,9,7,3,2) Year<-c(97,96,96,95,94,94,91,91,90,89,89,89,88,88,87,86,86,86,86,84,84,84,84,84, 83,83,82,81,81,80,80,79,79,79,78,78,77,76,76,75,75,75,75,74,74,73,73,73,72, 72,71,71,71,71,70,69,69,69,69,69,69,68,67,67,67,67,67) RollingStock<-c(2,2,3,2,1,1,1,1,2,3,1,1,1,2,1,2,1,3,2,2,1,2,2,3,1,2,1,1,2,3,1, 1,1,1,1,1,1,3,3,2,3,1,2,3,1,1,1,3,3,1,3,3,1,1,1,2,1,1,2,1,1,1,1,1,1,1,1) RollingStock <- as.factor(RollingStock) Traffic<-c(0.436,0.424,0.424,0.426,0.419,0.419,0.439,0.439,0.431,0.436,0.436, 0.436,0.443,0.443,0.397,0.414,0.414,0.414,0.414,0.389,0.389,0.389,0.389, 0.389,0.401,0.401,0.372,0.417,0.417,0.43,0.43,0.426,0.426,0.426,0.43,0.43, 0.425,0.426,0.426,0.436,0.436,0.436,0.436,0.452,0.452,0.433,0.433,0.433, 0.431,0.431,0.444,0.444,0.444,0.444,0.452,0.447,0.447,0.447,0.447,0.447, 0.447,0.449,0.459,0.459,0.459,0.459,0.459) Deaths<-c(7,1,1,1,5,2,4,2,1,1,2,5,35,1,4,1,2,1,1,3,1,3,13,2,1,1,1,4,1,2,1,5,7, 1,1,3,2,1,2,1,2,6,1,1,1,10,5,1,1,6,3,1,2,1,2,1,1,6,2,2,4,2,49,1,7,5,9) train2022 <- data.frame(Observation, Year, RollingStock, Traffic, Deaths) forsearch_glm(initial.sample = 100, response.cols = 5, indep.cols = 2:4, formula=Deaths~Year + RollingStock + Traffic, formula.cont.rhs="Year + Traffic", family = poisson("log"), data = train2022, n.obs.per.level = 1, estimate.phi = TRUE, skip.step1 = NULL, unblinded = TRUE, begin.diagnose=100)
Prepares summary statistics at each stage of forward search for subsequent plotting. Forward search is conducted in two steps: Step 1 to identify minimal set of observations to estimate unknown parameters, and Step 2 to add one observation at each stage such that observations in the set are best fitting at that stage.
forsearch_lm(formula, data, initial.sample=1000, n.obs.per.level = 1, skip.step1 = NULL, unblinded = TRUE, begin.diagnose = 100, verbose = TRUE)forsearch_lm(formula, data, initial.sample=1000, n.obs.per.level = 1, skip.step1 = NULL, unblinded = TRUE, begin.diagnose = 100, verbose = TRUE)
formula |
Fixed effects formula as described in help(lm). The only permitted operators are +, : , and * . Terms must be found in data or as constructed by I(xxx) where xxx is found in data |
data |
Name of database |
initial.sample |
Number of observations in Step 1 of forward search |
n.obs.per.level |
Number of observations per level of (possibly crossed) factor levels.Set to rank of X'X if model contains constructed variables such as I(x^3), for example in polynomial regression |
skip.step1 |
NULL or a vector of integers for observations to be included in Step 1 |
unblinded |
TRUE causes printing of presumed analysis structure |
begin.diagnose |
Numeric. Indicates where in code to begin printing diagnostics. 0 prints all; 100 prints none |
verbose |
TRUE causes function identifier to display before and after run |
Step 2 is determined by the results of Step 1, which itself is random. So, it is possible to reproduce the entire run by using the skip.step1 argument.
LIST
Rows in stage |
Observation numbers of rows included at each stage |
Standardized residuals |
Matrix of errors at each stage |
Number of model parameters |
Rank of model |
Sigma |
Estimate of random error at final stage; used to standardize all residuals |
Fixed parameter estimates |
Vector of parameter estimates at each stage |
s^2 |
Estimate of random error at each stage |
Leverage |
Matrix of leverage of each observation at each stage |
Modified Cook distance |
Estimate of sum of squared changes in parameter estimates at each stage |
Call |
Call to this function |
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
# Multiple regression Observation <- 1:16 y <- runif(16) x1 <- runif(16) x2 <- runif(16) x3 <- runif(16) lmtest1 <- data.frame(Observation,y,x1,x2,x3) forsearch_lm(formula=y~x1+x2+x3, data=lmtest1, initial.sample=200,begin.diagnose=100) ## Not run: # Analysis of variance Observation <- 1:30 y <- runif(30) AN1 <- as.factor(c(rep("A1",5),rep("A2",5),rep("A3",5))) AN1 <- c(AN1,AN1) AN2 <- as.factor(c(rep("B1",15),rep("B2",15))) lmtest2 <- data.frame(Observation,y,AN1,AN2) forsearch_lm(formula=y~AN1*AN2, data=lmtest2, initial.sample=200,begin.diagnose=100) # Analysis of covariance Observation <- 1:60 y <- runif(60) AN1 <- as.factor(c(rep("A1",10),rep("A2",10),rep("A3",10))) AN1 <- c(AN1,AN1) AN2 <- as.factor(c(rep("B1",30),rep("B2",30))) COV <- runif(60) lmtest3 <- data.frame(Observation,y,AN1,AN2,COV) forsearch_lm(formula=y~AN1*AN2+COV, data=lmtest3, initial.sample=200,begin.diagnose=100) # Polynomial regression C1 <- 7*runif(60) + 1 y <- 4 + C1 - 6*C1^2 + 9*C1^3 + rnorm(60) Observation <- 1:60 dfpoly <- data.frame(Observation,C1,y) forsearch_lm(formula = y ~ C1 + I(C1^2) + I(C1^3), data = dfpoly, initial.sample = 200, begin.diagnose=100) ## End(Not run)# Multiple regression Observation <- 1:16 y <- runif(16) x1 <- runif(16) x2 <- runif(16) x3 <- runif(16) lmtest1 <- data.frame(Observation,y,x1,x2,x3) forsearch_lm(formula=y~x1+x2+x3, data=lmtest1, initial.sample=200,begin.diagnose=100) ## Not run: # Analysis of variance Observation <- 1:30 y <- runif(30) AN1 <- as.factor(c(rep("A1",5),rep("A2",5),rep("A3",5))) AN1 <- c(AN1,AN1) AN2 <- as.factor(c(rep("B1",15),rep("B2",15))) lmtest2 <- data.frame(Observation,y,AN1,AN2) forsearch_lm(formula=y~AN1*AN2, data=lmtest2, initial.sample=200,begin.diagnose=100) # Analysis of covariance Observation <- 1:60 y <- runif(60) AN1 <- as.factor(c(rep("A1",10),rep("A2",10),rep("A3",10))) AN1 <- c(AN1,AN1) AN2 <- as.factor(c(rep("B1",30),rep("B2",30))) COV <- runif(60) lmtest3 <- data.frame(Observation,y,AN1,AN2,COV) forsearch_lm(formula=y~AN1*AN2+COV, data=lmtest3, initial.sample=200,begin.diagnose=100) # Polynomial regression C1 <- 7*runif(60) + 1 y <- 4 + C1 - 6*C1^2 + 9*C1^3 + rnorm(60) Observation <- 1:60 dfpoly <- data.frame(Observation,C1,y) forsearch_lm(formula = y ~ C1 + I(C1^2) + I(C1^3), data = dfpoly, initial.sample = 200, begin.diagnose=100) ## End(Not run)
Prepares summary statistics at each stage of forward search for subsequent plotting. Forward search is conducted in four steps: Step 0 to set up accounting for group structure, Step 1 to identify minimal set of observations to estimate unknown fixed parameters, Step 2 to identify the order of the remaining observations, and a final stage to extract the intermediate statistics based on increasing sample size.
forsearch_lme(fixedform, alldata, randomform, initial.sample=1000, n.obs.per.level=1, skip.step1=NULL, unblinded=TRUE, begin.diagnose = 100, verbose = TRUE)forsearch_lme(fixedform, alldata, randomform, initial.sample=1000, n.obs.per.level=1, skip.step1=NULL, unblinded=TRUE, begin.diagnose = 100, verbose = TRUE)
fixedform |
2-sided formula for fixed effects |
alldata |
data frame, first column of which must be "Observation" |
randomform |
1-sided formula for random effects |
initial.sample |
Number of observations in Step 1 of forward search |
n.obs.per.level |
Number of observations per level of (possibly crossed) factor levels |
skip.step1 |
NULL or a vector of integers for observations to be included in Step 1 |
unblinded |
TRUE causes printing of presumed analysis structure |
begin.diagnose |
Numeric indicator of place in coding to begin printing diagnostic information. 0 prints all information, 100 prints none. |
verbose |
TRUE causes function identifier to display before and after run |
data will be grouped within the function, regardless of initial layout. Step 2 is determined by the results of Step 1, which itself is random. So, it is possible to reproduce the entire run by using the skip.step1 argument. Variables in the randomform formula must be character variables, but *not* factors
LIST
Number of observations in Step 1 |
Number of observations included in Step 1 |
Step 1 observation numbers |
Observation numbers useful in skipping step 1 |
Rows by outer subgroup |
List of row numbers, by outer subgroup |
Rows by outer-inner subgroups |
List of row numbers, by outer-inner subgroup |
Rows in stage |
Observation numbers of rows included at each stage |
Sigma |
Estimate of random error at final stage; used to standardize all residuals |
Standardized residuals |
Matrix of errors at each stage |
Fixed parameter estimates |
Matrix of parameter estimates at each stage |
Random parameter estimates |
Matrix of parameter estimates at each stage |
Leverage |
Matrix of leverage of each observation at each stage |
Modified Cook distance |
Estimate of sum of squared changes in parameter estimates at each stage |
Dims |
Dims from fit of lme function |
t statistics |
t statistics for each fixed parameter |
Fit statistics |
AIC, BIC, and log likelihood |
Call |
Call to this function |
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000. Pinheiro, JC and DM Bates. Mixed-Effects Models in S and S-Plus, Springer, New York, 2000. https://CRAN.R-project.org/package=nlme
# Multiple regression in grouped data Observation <- 1:16 y <- runif(16) x1 <- runif(16) x2 <- runif(16) x3 <- runif(16) group <- as.factor(rep(c("G1","G2"),each=8)) lmetest1 <- data.frame(Observation,y,x1,x2,x3,group) forsearch_lme(fixedform=y~x1+x2+x3, alldata=lmetest1, randomform= ~1|group, n.obs.per.level=1, initial.sample=200) ## Not run: # Analysis of variance in grouped data Observation <- 1:60 y <- runif(60) AN1 <- as.factor(c(rep("A1",5),rep("A2",5),rep("A3",5))) AN1 <- c(AN1,AN1,AN1,AN1) AN2 <- as.factor(c(rep("B1",15),rep("B2",15))) AN2 <- c(AN2,AN2) group <- as.factor(rep(c("G1","G2"),each=30)) lmetest2 <- data.frame(Observation,y,AN1,AN2,group) forsearch_lme(fixedform=y~AN1*AN2, alldata=lmetest2, randomform= ~1|group, initial.sample=500) # Analysis of covariance in grouped data Observation <- 1:120 y <- runif(120) AN1 <- as.factor(c(rep("A1",10),rep("A2",10),rep("A3",10))) AN1 <- c(AN1,AN1,AN1,AN1) AN2 <- as.factor(c(rep("B1",10),rep("B2",10))) AN2 <- c(AN2,AN2,AN2,AN2,AN2,AN2) COV <- runif(120) group <- as.factor(rep(c("G1","G2"),each=30)) group <- c(group,group) lmetest3 <- data.frame(Observation,y,AN1,AN2,COV,group) forsearch_lme(fixedform=y~AN1*AN2+COV, alldata=lmetest3, randomform= ~ 1 | group, initial.sample=500) ## End(Not run)# Multiple regression in grouped data Observation <- 1:16 y <- runif(16) x1 <- runif(16) x2 <- runif(16) x3 <- runif(16) group <- as.factor(rep(c("G1","G2"),each=8)) lmetest1 <- data.frame(Observation,y,x1,x2,x3,group) forsearch_lme(fixedform=y~x1+x2+x3, alldata=lmetest1, randomform= ~1|group, n.obs.per.level=1, initial.sample=200) ## Not run: # Analysis of variance in grouped data Observation <- 1:60 y <- runif(60) AN1 <- as.factor(c(rep("A1",5),rep("A2",5),rep("A3",5))) AN1 <- c(AN1,AN1,AN1,AN1) AN2 <- as.factor(c(rep("B1",15),rep("B2",15))) AN2 <- c(AN2,AN2) group <- as.factor(rep(c("G1","G2"),each=30)) lmetest2 <- data.frame(Observation,y,AN1,AN2,group) forsearch_lme(fixedform=y~AN1*AN2, alldata=lmetest2, randomform= ~1|group, initial.sample=500) # Analysis of covariance in grouped data Observation <- 1:120 y <- runif(120) AN1 <- as.factor(c(rep("A1",10),rep("A2",10),rep("A3",10))) AN1 <- c(AN1,AN1,AN1,AN1) AN2 <- as.factor(c(rep("B1",10),rep("B2",10))) AN2 <- c(AN2,AN2,AN2,AN2,AN2,AN2) COV <- runif(120) group <- as.factor(rep(c("G1","G2"),each=30)) group <- c(group,group) lmetest3 <- data.frame(Observation,y,AN1,AN2,COV,group) forsearch_lme(fixedform=y~AN1*AN2+COV, alldata=lmetest3, randomform= ~ 1 | group, initial.sample=500) ## End(Not run)
Prepares summary statistics at each stage of forward search for subsequent plotting. Forward search is conducted in two steps: Step 1 to identify minimal set of observations to estimate unknown parameters, and Step 2 to add one observation at each stage such that observations in the set are best fitting at that stage.
forsearch_nls(phaselist, data, poolstart, poolformula, algorithm= "default", controlarg=NULL, initial.sample=1000, skip.step1=NULL, begin.diagnose=100, verbose=TRUE)forsearch_nls(phaselist, data, poolstart, poolformula, algorithm= "default", controlarg=NULL, initial.sample=1000, skip.step1=NULL, begin.diagnose=100, verbose=TRUE)
phaselist |
LIST of formula, formulacont, start, nopl for each phase |
data |
Name of database. First 2 variables are Observation and Phases (both mandatory) |
poolstart |
List Start values for Step 2 |
poolformula |
Formula for pooled data from all phases for Step 2 |
algorithm |
algorithm for nls function. |
controlarg |
nls control. Default is NULL to use preset nls.control |
initial.sample |
Number of observation sets in Step 1 of forward search |
skip.step1 |
NULL or a vector of integers for observations to be included in Step 1 |
begin.diagnose |
Numeric. Indicates where in code to begin printing diagnostics. 0 prints all; 100 prints none |
verbose |
TRUE causes function identifier to display before and after run |
All datasets are considered to be in phases. See vignette for definition and discussion. There is a phaselist for each phase and an element for each phaselist input variable. In addition, there is a (pool)start and a (pool)formula input variable for the pooled dataset.
LIST
Rows in stage |
Observation numbers of rows included at each stage |
Standardized residuals |
Matrix of errors at each stage |
Number of model parameters |
Same as number of levels of poolstart input variable |
Sigma |
Estimate of random error at final stage; used to standardize all residuals |
Fixed parameter estimates |
Vector of parameter estimates at each stage |
s^2 |
Estimate of random error at each stage |
Call |
Call to this function |
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000. Pinheiro, JC and DM Bates. Mixed Effects Models in S and S-PLUS, Springer, New York, 2000. Example from nlstools package
## Not run: t<-(0:35)/3 VO2<-c(377.1111,333.3333,352.1429,328.7500,369.8750,394.4000,352.6667,337.3333, 366.4286,364.0000,293.8889,387.0000,364.8889,342.2222,400.3000,375.1111, 320.5556,385.1667,527.0714,688.6364,890.8182,1145.1538,1254.9091,1327.5000, 1463.9000,1487.8333,1586.6667,1619.1000,1494.4167,1640.4545,1643.3750, 1583.6364,1610.8000,1568.5000,1464.5833,1652.8000) Observation <- 1:36 Phases <- as.factor(c(rep("REST",18), rep("EXERCISE",18))) test01 <- data.frame(Observation,Phases,t,VO2) formula.1 <-as.formula(VO2~VO2rest) formulacont.1 <- as.formula(VO2~VO2rest) start.1 <- list(VO2rest = 400) nopl.1 <- 1 formula.2<- as.formula(VO2~(VO2rest+(VO2peak-VO2rest)*(1-exp(-(t-5.883)*I(1/mu))))) formulacont.2<- as.formula(VO2~(VO2rest+(VO2peak-VO2rest)*(1-exp(-(t-5.883)*I(1/mu))))) start.2 <- list(VO2rest = 400, VO2peak = 1600, mu = 1) nopl.2 <- 6 phaselist <- list( REST= list(formula=formula.1,formulacont=formulacont.1,start=start.1,nopp=nopl.1), EXERCISE= list(formula=formula.2,formulacont=formulacont.2,start=start.2,nopp=nopl.2)) pstart <- list(VO2rest=400, VO2peak = 1600, mu = 1) pformula <- as.formula(VO2~(t<=5.883)*(VO2rest)+ (t>5.883)*(VO2rest+(VO2peak-VO2rest)* (1-exp(-(t-5.883)*I(1/mu))))) forsearch_nls(phaselist=phaselist, data=test01, poolstart=pstart, poolformula=pformula, algorithm="default", controlarg=nls.control(maxiter=50,warnOnly=TRUE), initial.sample = 155) ## End(Not run)## Not run: t<-(0:35)/3 VO2<-c(377.1111,333.3333,352.1429,328.7500,369.8750,394.4000,352.6667,337.3333, 366.4286,364.0000,293.8889,387.0000,364.8889,342.2222,400.3000,375.1111, 320.5556,385.1667,527.0714,688.6364,890.8182,1145.1538,1254.9091,1327.5000, 1463.9000,1487.8333,1586.6667,1619.1000,1494.4167,1640.4545,1643.3750, 1583.6364,1610.8000,1568.5000,1464.5833,1652.8000) Observation <- 1:36 Phases <- as.factor(c(rep("REST",18), rep("EXERCISE",18))) test01 <- data.frame(Observation,Phases,t,VO2) formula.1 <-as.formula(VO2~VO2rest) formulacont.1 <- as.formula(VO2~VO2rest) start.1 <- list(VO2rest = 400) nopl.1 <- 1 formula.2<- as.formula(VO2~(VO2rest+(VO2peak-VO2rest)*(1-exp(-(t-5.883)*I(1/mu))))) formulacont.2<- as.formula(VO2~(VO2rest+(VO2peak-VO2rest)*(1-exp(-(t-5.883)*I(1/mu))))) start.2 <- list(VO2rest = 400, VO2peak = 1600, mu = 1) nopl.2 <- 6 phaselist <- list( REST= list(formula=formula.1,formulacont=formulacont.1,start=start.1,nopp=nopl.1), EXERCISE= list(formula=formula.2,formulacont=formulacont.2,start=start.2,nopp=nopl.2)) pstart <- list(VO2rest=400, VO2peak = 1600, mu = 1) pformula <- as.formula(VO2~(t<=5.883)*(VO2rest)+ (t>5.883)*(VO2rest+(VO2peak-VO2rest)* (1-exp(-(t-5.883)*I(1/mu))))) forsearch_nls(phaselist=phaselist, data=test01, poolstart=pstart, poolformula=pformula, algorithm="default", controlarg=nls.control(maxiter=50,warnOnly=TRUE), initial.sample = 155) ## End(Not run)
Runs the defined, grouped linear mixed effects (lme) model. Displays the resulting fixed and random coefficients. Attaches codes for identifying them to the plotting functions of this package.
identifyCoeffs(fixed, data, random, XmaxIter = 1000, XmsMaxIter = 1000, Xtolerance = 0.01, XniterEM = 1000, XmsMaxEval = 400, XmsTol = 1e-05, Xopt = "optim", verbose = TRUE)identifyCoeffs(fixed, data, random, XmaxIter = 1000, XmsMaxIter = 1000, Xtolerance = 0.01, XniterEM = 1000, XmsMaxEval = 400, XmsTol = 1e-05, Xopt = "optim", verbose = TRUE)
fixed |
2-sided formula for fixed effects |
data |
Name of file (to be) run by forsearch_lme |
random |
1-sided formula for random effects |
XmaxIter |
lme control parameter |
XmsMaxIter |
lme control parameter |
Xtolerance |
lme control parameter |
XniterEM |
lme control parameter |
XmsMaxEval |
lme control parameter |
XmsTol |
lme control parameter |
Xopt |
lme control parameter |
verbose |
If TRUE, indicates beginning and end of function |
Plotting functions cannot plot more than a few coefficients on one graph. This function prepares an index of the coefficients so that the user can more easily identify which ones should appear together in a plot.
Index of fixed and random coefficients from forsearch_lme.
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
info3 <- system.file("extdata","Machines.O.R",package="forsearch"); info3 <- source(info3); info3 <- info3[[1]]; identifyCoeffs(fixed=score~1, data=info3, random= ~1 | Worker)info3 <- system.file("extdata","Machines.O.R",package="forsearch"); info3 <- source(info3); info3 <- info3[[1]]; identifyCoeffs(fixed=score~1, data=info3, random= ~1 | Worker)
Runs the defined linear (lm) model. Displays the resulting coefficients. Attaches codes for identifying them to the plotting functions of this package.
identifyFixedCoeffs(formula, data, verbose = TRUE)identifyFixedCoeffs(formula, data, verbose = TRUE)
formula |
2-sided formula for fixed effects |
data |
Name of file (to be) run by forsearch_lm |
verbose |
If TRUE, indicates beginning and end of function |
Plotting functions cannot plot more than a few coefficients on one graph. This function prepares an index of the coefficients so that the user can more easily identify which ones should appear together in a plot.
Index of coefficients from forsearch_lm.
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
info3 <- system.file("extdata", "crossdata.R", package="forsearch"); crossdata <- source(info3); crossdata <- crossdata[[1]]; identifyFixedCoeffs(formula=y~x1*x2, data=crossdata)info3 <- system.file("extdata", "crossdata.R", package="forsearch"); crossdata <- source(info3); crossdata <- crossdata[[1]]; identifyFixedCoeffs(formula=y~x1*x2, data=crossdata)
Restricts Step 1 of forward search procedures to ensure that every possible combination of levels is included for observations containing factors
picksome(subsetlist, nobs, initial.sample, n.obs.per.level, rank)picksome(subsetlist, nobs, initial.sample, n.obs.per.level, rank)
subsetlist |
List, each element is a data frame of 2 columns with code indicating the highest possible level of interaction to which each observation can belong. Usually, output from variablelist function. |
nobs |
Number of observations in data frame containing observations of forward search. |
initial.sample |
Number of randomized sets of observations in Step 1 of forward search. |
n.obs.per.level |
Number of observations to pull from each level. |
rank |
Rank of the X matrix of the analytical function to be used on data frame. |
Support function, usually not called independently. Argument n.obs.per.level is set by user in forsearch_xxx function call.
Matrix, each row of which identifies observations in each set of random sample of observations.
William R. Fairweather
Plot output from forsearch_glm to show change in AIC statistics as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.AICX(forn, maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption title here", wmf = "Put_plot_file_title_here", Cairo=TRUE, printgraph=TRUE,addline="none", verbose = TRUE)plotdiag.AICX(forn, maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption title here", wmf = "Put_plot_file_title_here", Cairo=TRUE, printgraph=TRUE,addline="none", verbose = TRUE)
forn |
Name of output file from forsearch_glm |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
addline |
add a line to the graph; "none", "loess", or "straight"); abbreviation allowed |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot AIC statistics from forsearch_glm
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Executes all the plotting functions for a select analytical function such as lm or glm; default omits titles and subtitles and attempts to plot all fixed and random coefficients.
plotdiag.allgraphs(object, mt=" ", st=" ", cpt=" ", blind.label=FALSE, cc=NULL, ccrand = NULL,Cairo=TRUE)plotdiag.allgraphs(object, mt=" ", st=" ", cpt=" ", blind.label=FALSE, cc=NULL, ccrand = NULL,Cairo=TRUE)
object |
Name of forsearch object file |
mt |
Maintitle of graph |
st |
Subtitle of graph |
cpt |
Caption on the graph |
blind.label |
TRUE causes 'blind' to be added to graph and to file name for fixed parameters |
cc |
Fixed variable code numbers of coefficients to be included in graph |
ccrand |
Random variable code numbers of parameters to be included in graph |
Cairo |
TRUE causes use of Cairo graphics |
Prints search history and creates graphical files in current subdirectory
William R. Fairweather
Plot output from forsearch_xxx to show change in anova p-values as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.ANOX2(forn, anova.rows=NULL, ylab.extend=c("proportionality","variance"), maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here",wmf = "Put_stored_name_here", Cairo=TRUE, printgraph=TRUE,legend = "Dummy legend name", verbose = TRUE)plotdiag.ANOX2(forn, anova.rows=NULL, ylab.extend=c("proportionality","variance"), maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here",wmf = "Put_stored_name_here", Cairo=TRUE, printgraph=TRUE,legend = "Dummy legend name", verbose = TRUE)
forn |
Name of output file from forsearch_xxx |
anova.rows |
Row numbers of p values to include together on the plot |
ylab.extend |
Type of anova table. "proportionality" is a test of proportionality for a coxph analysis; "variance" is a test of null hypothesis of a lm or lme test |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
legend |
Name of legend |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot anova test p values from forsearch_lm or forsearch_lme
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_xxx to show change in fixed coefficients as the number of observations in the forward search procedure increases. Save plot in folder containing working directory. Run on blinded data only.
plotdiag.blind.fixed(forn, coeff.codenums=NULL, maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here",wmf = "Put_stored_name_here", Cairo=TRUE, printgraph=TRUE,legend = "Dummy legend name", verbose = TRUE)plotdiag.blind.fixed(forn, coeff.codenums=NULL, maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here",wmf = "Put_stored_name_here", Cairo=TRUE, printgraph=TRUE,legend = "Dummy legend name", verbose = TRUE)
forn |
Name of output file from forsearch_xxx |
coeff.codenums |
Numeric vector of coefficients to include together on the plot. Codes are output by identifyFixedCoeffs (for lm files) or by identifyCoeffs function (for lme files) |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
legend |
Name of legend |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot fixed coefficient statistics from forsearch_lm or forsearch_lme
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_lm or forsearch_lme to show change in Modified Cook's distance as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.Cook(forn, maintitle = "Put main title here", subtitle = "Put subtitle here", caption = "Put caption here", wmf = "Put_plot_file_title_here", Cairo=TRUE, printgraph=TRUE, addline = "none", verbose = TRUE)plotdiag.Cook(forn, maintitle = "Put main title here", subtitle = "Put subtitle here", caption = "Put caption here", wmf = "Put_plot_file_title_here", Cairo=TRUE, printgraph=TRUE, addline = "none", verbose = TRUE)
forn |
Name of forward search output file |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
addline |
Character variable to add a line to the graph; options: "none", "loess", and "straight"; abbreviation allowed |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot Cook distance statistics from forsearch_lm or forsearch_lme
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_glm to show change in deviance residuals or augmented deviance residuals, either of which can be squared, as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.deviance.residuals(forn, squared = FALSE, augmented=TRUE, hilos = c(1, 0), maintitle="Put main title here", subtitle="Put subtitle here", caption="Put caption here", wmf= "Put_graph_title_here", Cairo=TRUE,printgraph=TRUE, legend = "Dummy legend name", verbose = TRUE)plotdiag.deviance.residuals(forn, squared = FALSE, augmented=TRUE, hilos = c(1, 0), maintitle="Put main title here", subtitle="Put subtitle here", caption="Put caption here", wmf= "Put_graph_title_here", Cairo=TRUE,printgraph=TRUE, legend = "Dummy legend name", verbose = TRUE)
forn |
Name of forward search output file |
squared |
TRUE causes residuals to be squared before plotting |
augmented |
TRUE causes graphing of augmented deviance residuals, see Details |
hilos |
Number of observations having high and number having low values of residuals to identify. No low values are identified for squared residual plot |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Caption of plot |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
legend |
Legend title |
verbose |
If TRUE, indicates beginning and end of function |
We reserve the use of the term 'Deviance residuals' to deviance residuals of the observations that were used to create the model fit, and use the term 'Augmented deviance residuals' to refer to deviance residuals of all available observations. The latter are created by predicting the fit of the model to all observations.
Process and plot changes in deviance residuals or squared deviance residuals from forsearch_glm
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_glm to show change in deviances as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.deviances(forn, devtype, maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here", wmf = "Put_plot_file_title_here", Cairo=TRUE, printgraph=TRUE,addline="none", verbose = TRUE)plotdiag.deviances(forn, devtype, maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here", wmf = "Put_plot_file_title_here", Cairo=TRUE, printgraph=TRUE,addline="none", verbose = TRUE)
forn |
Name of output file from forsearch_glm |
devtype |
Type of deviance: "R" or "N" for Residual deviance or Null deviance |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
addline |
add a line to the graph; abbreviation allowed; "none","loess", or "straight" |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot deviances from forsearch_glm
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_lme to show change in AIC, BIC, and log likelihood as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.fit3(forn, maintitle = "Put main title here", subtitle = "Put subtitle here", caption = "Put caption here", wmf = "Put_stored_name_here", Cairo=TRUE,printgraph=TRUE, legend="Dummy legend name", verbose = TRUE)plotdiag.fit3(forn, maintitle = "Put main title here", subtitle = "Put subtitle here", caption = "Put caption here", wmf = "Put_stored_name_here", Cairo=TRUE,printgraph=TRUE, legend="Dummy legend name", verbose = TRUE)
forn |
Name of output file from forsearch_lm |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
legend |
Legend name |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot trends of AIC, BIC, and log likelihood statistics from forsearch_lme
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_lm or forsearch_lme to show change in leverage of each observation as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.leverage(forn, hilos = c(1, 0), maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here",wmf = "Put_graph_title_here", Cairo=TRUE, printgraph = TRUE, verbose = TRUE)plotdiag.leverage(forn, hilos = c(1, 0), maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here",wmf = "Put_graph_title_here", Cairo=TRUE, printgraph = TRUE, verbose = TRUE)
forn |
Name of forward search output file |
hilos |
Vector with number of highest observations and number of lowest observations on graph to identify |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot Cook distance statistics from forsearch_lm or forsearch_lme
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_cph to show change in loglik pairs as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.loglik(forn, maintitle= "Put main title here", subtitle= "Put subtitle here" , caption="Put caption here", wmf = "Put_stored_name_here", Cairo=TRUE, printgraph = TRUE, verbose=TRUE)plotdiag.loglik(forn, maintitle= "Put main title here", subtitle= "Put subtitle here" , caption="Put caption here", wmf = "Put_stored_name_here", Cairo=TRUE, printgraph = TRUE, verbose=TRUE)
forn |
Name of output file from forsearch_cph |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot Wald Test statistics from forsearch_cph
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_cph to show change in likelihood ratio test as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.lrt(forn, maintitle= "Put main title here", subtitle= "Put subtitle here" , caption="Put caption here", wmf = "Put_graph_filename_here", Cairo=TRUE, printgraph = TRUE, addline=c("none","loess","straight"), verbose=TRUE)plotdiag.lrt(forn, maintitle= "Put main title here", subtitle= "Put subtitle here" , caption="Put caption here", wmf = "Put_graph_filename_here", Cairo=TRUE, printgraph = TRUE, addline=c("none","loess","straight"), verbose=TRUE)
forn |
Name of output file from forsearch_cph |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
addline |
Add a line to the graph; abbreviation allowed. Default none |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot likelihood ratio test statistics from forsearch_cph
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_xxx to show change in fixed coefficients as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.params.fixed(forn, coeff.codenums=NULL, maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here",wmf = "Put_stored_name_here", Cairo=TRUE, printgraph=TRUE,legend = "Dummy legend name", verbose = TRUE)plotdiag.params.fixed(forn, coeff.codenums=NULL, maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here",wmf = "Put_stored_name_here", Cairo=TRUE, printgraph=TRUE,legend = "Dummy legend name", verbose = TRUE)
forn |
Name of output file from forsearch_xxx |
coeff.codenums |
Numeric vector of coefficients to include together on the plot. Codes are output by identifyFixedCoeffs (for lm files) or by identifyCoeffs function (for lme files) |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
legend |
Name of legend |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot fixed coefficient statistics from forsearch_lm or forsearch_lme
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_lme to show change in root mean squares of random coefficients as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.params.random(forn, coeff.codenums=NULL, asfacets=FALSE, facetdir=c("h","v"), maintitle = "Put maintitle here", subtitle = "Put subtitle here", caption = "Put caption here", wmf = "Put_stored_name_here", Cairo=TRUE, printgraph = TRUE, legend = "Dummy legend name", verbose = TRUE)plotdiag.params.random(forn, coeff.codenums=NULL, asfacets=FALSE, facetdir=c("h","v"), maintitle = "Put maintitle here", subtitle = "Put subtitle here", caption = "Put caption here", wmf = "Put_stored_name_here", Cairo=TRUE, printgraph = TRUE, legend = "Dummy legend name", verbose = TRUE)
forn |
Name of output file from forsearch_lme |
coeff.codenums |
columns of output file to be included in graph |
asfacets |
TRUE causes printing in facets |
facetdir |
"v" lays out the facets vertically, "h" lays them out horizontally |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
legend |
Name of legend |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot RMS of random coefficients from forsearch_lme
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_glm to show change in phiHat statistics as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.phihatx(forn, maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here", wmf = "Put_graph_filename_here", Cairo=TRUE, printgraph=TRUE, addline="none", verbose = TRUE)plotdiag.phihatx(forn, maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here", wmf = "Put_graph_filename_here", Cairo=TRUE, printgraph=TRUE, addline="none", verbose = TRUE)
forn |
Name of output file from forsearch_glm |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
addline |
add a line to the graph; abbreviation allowed; "none", "loess", or "straight"" |
printgraph |
TRUE causes graph to print to file and closes device |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot phiHat statistics from forsearch_glm
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_lm or forsearch_lme to show change in residuals or squared residuals as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.residuals(forn, squared = FALSE, hilos = c(1, 0), maintitle, subtitle, caption, wmf, Cairo=TRUE,printgraph=TRUE, legend = "Dummy legend name", verbose = TRUE)plotdiag.residuals(forn, squared = FALSE, hilos = c(1, 0), maintitle, subtitle, caption, wmf, Cairo=TRUE,printgraph=TRUE, legend = "Dummy legend name", verbose = TRUE)
forn |
Name of forward search output file |
squared |
TRUE causes residuals to be squared before plotting |
hilos |
Number of observations having high and number having low values of residuals to identify. No low values are identified for squared residual plot. |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Caption of plot |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
legend |
Legend title |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot changes in residuals or squared residuals from forsearch_lm or forsearch_lme
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_lm to show change in residual variation as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.s2(forn, maintitle = "Put main title here", subtitle = "Put subtitle here", caption = "Put caption here", wmf = "Put_graph_filename_here", Cairo=TRUE,printgraph=TRUE, addline = c("none","loess","straight"), verbose = TRUE)plotdiag.s2(forn, maintitle = "Put main title here", subtitle = "Put subtitle here", caption = "Put caption here", wmf = "Put_graph_filename_here", Cairo=TRUE,printgraph=TRUE, addline = c("none","loess","straight"), verbose = TRUE)
forn |
Name of output file from forsearch_lm |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
addline |
add a line to the graph; abbreviation allowed |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot residual variation statistics from forsearch_lm
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_lm or forsearch_lme to show change in t statistics as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.tstats(forn, coeff.codenums=NULL, maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here", wmf = "Put_stored_name_here", Cairo=TRUE, printgraph=TRUE,legend = "Dummy legend name", verbose = TRUE)plotdiag.tstats(forn, coeff.codenums=NULL, maintitle = "Put main title here", subtitle = "Put subtitle here", caption="Put caption here", wmf = "Put_stored_name_here", Cairo=TRUE, printgraph=TRUE,legend = "Dummy legend name", verbose = TRUE)
forn |
Name of output file from forsearch_lm or forsearch_lme |
coeff.codenums |
Numeric vector of coefficients to include together on the plot. Codes are output by identifyFixedCoeffs (for lm files) or by identifyCoeffs function (for lme files) |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
legend |
Name of legend |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot t statistics of fixed coefficients from forsearch_lm or forsearch_lme
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
Plot output from forsearch_cph to show change in Wald test as the number of observations in the forward search procedure increases. Save plot in folder containing working directory.
plotdiag.Wald(forn, maintitle= "Put main title here", subtitle= "Put subtitle here" , caption="Put caption here", wmf = "Put_graph_filename_here", Cairo=TRUE, printgraph = TRUE, addline=c("none","loess","straight"), verbose=TRUE)plotdiag.Wald(forn, maintitle= "Put main title here", subtitle= "Put subtitle here" , caption="Put caption here", wmf = "Put_graph_filename_here", Cairo=TRUE, printgraph = TRUE, addline=c("none","loess","straight"), verbose=TRUE)
forn |
Name of output file from forsearch_cph |
maintitle |
Main title of plot |
subtitle |
Subtitle of plot |
caption |
Content of caption |
wmf |
File name of stored plot; omit ".wmf" |
Cairo |
TRUE causes use of Cairo graphics |
printgraph |
TRUE causes graph to print to file and closes device |
addline |
Add a line to the graph; abbreviation allowed. Default none |
verbose |
If TRUE, indicates beginning and end of function |
Process and plot Wald Test statistics from forsearch_cph
William R. Fairweather
Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000.
The forward search functions output a list of vectors, each of which indicates which observations are in the model at each stage of the search. This function processes that list to create a more easily understood matrix of the observation numbers that are newly entered into the model and any that were temporarily removed from the model over the course of the search.
search.history(list1, verbose = TRUE)search.history(list1, verbose = TRUE)
list1 |
Name of a forsearch_xxx output file |
verbose |
If TRUE, indicates beginning and end of function |
Printout of matrix showing evolution of observations to enter or leave the model during the course of the forward search
William R. Fairweather
info3 <- system.file("extdata", "crossdata.for1.R", package="forsearch"); info3 <- source(info3); info3 <- info3[[1]]; search.history(list1=info3, verbose=TRUE)info3 <- system.file("extdata", "crossdata.for1.R", package="forsearch"); info3 <- source(info3); info3 <- info3[[1]]; search.history(list1=info3, verbose=TRUE)
Output of forsearch_xxx function can be voluminous. This function displays the output in an abbreviated format. Primarily for programmer use.
showme(x, verbose = TRUE)showme(x, verbose = TRUE)
x |
Name of forsearch_xxx output file |
verbose |
If TRUE, indicates the beginning and end of function run |
Abbreviated printout of output of forsearch_lm function
William R. Fairweather
For a data frame with factor variables V1, V2, V3, etc having levels n1, n2, n3, etc, lists the n1*n2*n3*... possible interaction levels and identifies which of the observations of the data frame belong in which of these interaction levels.
variablelist(datadf, prank)variablelist(datadf, prank)
datadf |
Data frame of independent variables in analysis. First column of data frame is Observation number |
prank |
Number of continuous variables among independent variables |
Support function, usually not called independently
List, each element is a data frame of 2 columns with code indicating the highest possible level of interaction to which each observation can belong
William R. Fairweather