Table 1 Abbreviation and explanation

OE

Orthogonally encoded features

FI-BOTH

Frequent itemsets are extracted from both cleaved and non-cleaved instances as features

FI-YES

Frequent itemsets are extracted only from cleaved instances as features

FI-NO

Frequent itemsets are extracted only from non-cleaved instances as features

SUP-m

Frequent itemset minimum support threshold m as percentage is used; if it is not present then 3% is used

PCA-100

Principal component analysis is used for feature selection and 100 features are selected

FI-CENTER

Frequent itemsets which have items in \(P_1\) or \(P^{\prime }_1\) position are selected as features

uni

ANOVA F value’s used for feature selection

RFE

Recursive feature elimination is used for feature selection

SNA-100

Social network analysis is used for feature selection and 100 features are selected