Package 'rCBA'

Build classifier function (Apriori-based)

Description

Automatic build of the classification model using the Apriori algorithm from the arules

Usage

build(trainData, className = NA, pruning = TRUE, sa = list(),
  verbose = TRUE, parallel = TRUE)

Arguments

trainData	data.frame or transactions from arules with input data
className	column name with the target class - default is the last column
pruning	performing pruning while building the model
sa	simulated annealing setting. Default values: list(temp=100.0, alpha=0.05, tabuRuleLength=5, timeout=10)
verbose	verbose indicator
parallel	parallel indicator

Value

list with parameters and model as data.frame with rules

Examples

library("rCBA")
data("iris")

output <- rCBA::build(iris,sa = list(alpha=0.5), parallel=FALSE) # speeding up the cooling
model <- output$model

predictions <- rCBA::classification(iris, model)
table(predictions)
sum(as.character(iris$Species)==as.character(predictions), na.rm=TRUE) / length(predictions)

---

Build classifier function (FP-Growth-based)

Description

Automatic build of the classification model using the FP-Growth algorithm

Usage

buildFPGrowth(train, className = NULL, verbose = TRUE,
  parallel = TRUE)

Arguments

train	data.frame or transactions from arules with input data
className	column name with the target class - default is the last column
verbose	verbose indicator
parallel	parallel indicator

Value

list with parameters and model as data.frame with rules

Examples

library("rCBA")
data("iris")

output <- rCBA::buildFPGrowth(iris[sample(nrow(iris), 10),], "Species",
 parallel=FALSE, verbose=TRUE)
inspect(output$model)

---

A classification function

Description

A classification function

Usage

classification(test, rules, verbose = TRUE)

Arguments

test	data.frame or transactions from arules with input data
rules	data.frame with rules
verbose	verbose indicator

Value

vector with classifications

Examples

library("arules")
library("rCBA")
data("iris")

train <- sapply(iris, as.factor)
train <- data.frame(train, check.names=FALSE)
txns <- as(train,"transactions")

rules = apriori(txns, parameter=list(support=0.03, confidence=0.03, minlen=2),
appearance = list(rhs=c("Species=setosa", "Species=versicolor", "Species=virginica"),default="lhs"))

predictions <- rCBA::classification(train,rules)
table(predictions)
sum(as.character(train$Species)==as.character(predictions),na.rm=TRUE)/length(predictions)

---

FP-Growth

Description

FP-Growth algorithm - Jiawei Han, Jian Pei, and Yiwen Yin. Mining frequent patterns without candidate generation. SIGMOD Rec. 29, 2 (2000) <doi:10.1145/335191.335372>

Usage

fpgrowth(train, support = 0.01, confidence = 1, maxLength = 5,
  consequent = NULL, verbose = TRUE, parallel = TRUE)

Arguments

train	data.frame or transactions from arules with input data
support	minimum support
confidence	minimum confidence
maxLength	maximum length
consequent	filter consequent - column name with consequent/target class
verbose	verbose indicator
parallel	parallel indicator

Examples

library("rCBA")
data("iris")

train <- sapply(iris,as.factor)
train <- data.frame(train, check.names=FALSE)
txns <- as(train,"transactions")

rules = rCBA::fpgrowth(txns, support=0.03, confidence=0.03, maxLength=2, consequent="Species",
           parallel=FALSE)

predictions <- rCBA::classification(train,rules)
table(predictions)
sum(as.character(train$Species)==as.character(predictions),na.rm=TRUE)/length(predictions)

prunedRules <- rCBA::pruning(train, rules, method="m2cba", parallel=FALSE)
predictions <- rCBA::classification(train, prunedRules)
table(predictions)
sum(as.character(train$Species)==as.character(predictions),na.rm=TRUE)/length(predictions)

---

Conversion of data.frame to rules from arules

Description

Conversion of data.frame to rules from arules

Usage

frameToRules(model)

Arguments

model

data.frame with rules

Value

arules rules representation

Examples

library("rCBA")

model <- data.frame("rules" = c("{X=1} => {Y=1}","{X=0} => {Y=0}"),
                    "support" = c(0.5,0.5),
                    "confidence" = c(0.5,0.5),
                    "lift" = c(1.0,1.0))

rules <- rCBA::frameToRules(model)
inspect(rules)

---

A Pruning function

Description

A Pruning function

Usage

pruning(train, rules, method = "m2cba", verbose = TRUE,
  parallel = TRUE)

Arguments

train	trainData data.frame or transactions from arules with input data
rules	data.frame with rules
method	pruning method m2cba(default)|m1cba|dcbrcba
verbose	verbose indicator
parallel	parallel indicator

Value

data.frame with pruned rules

Examples

library("arules")
library("rCBA")
data("iris")

train <- sapply(iris,as.factor)
train <- data.frame(train, check.names=FALSE)
txns <- as(train,"transactions")

rules = apriori(txns, parameter=list(support=0.03, confidence=0.03, minlen=2),
appearance = list(rhs=c("Species=setosa", "Species=versicolor", "Species=virginica"),default="lhs"))

print(length(rules))
prunedRules <- rCBA::pruning(train, rules, method="m2cba", parallel=FALSE)
print(length(prunedRules))

---