Empirical study of seven data mining algorithms on different characteristics of datasets for biomedical classification applications

Table 1 Profile of research data sets

Name of dataset	Sample size	Number of attributes	Missing values?	Task	Area
Iris	150	4	No	Multi-class	Life
Adult^a	32,561	13	Yes	Binary-class	Social
Wine	178	13	No	Multi-class	Physical
Car evaluation	1728	6	No	Multi-class	–
Breast cancer Wisconsin^a	699	9	Yes	Binary-class	Life
Wdbc^a	569	30	No	Binary-class	Life
Wpbc^a	198	31	Yes	Binary-class	Life
Abalone	4177	8	No	Multi-class	Life
Wine quality_red^a	1599	11	No	Multi-class	Business
Wine quality_white^a	4898	11	No	Multi-class	Business
Heart disease^a	303	13	Yes	Multi-class	Life
Poker hand^a	25,010	10	No	Multi-class	Game

^aThe dataset ‘Adult’ is a subset of the database ‘Adult Data Set’. The datasets ‘Breast cancer Wisconsin’, ‘Wdbc’ and ‘Wpbc’ are three subsets come from the same database ‘Breast Cancer Wisconsin (diagnostic) data set’. The datasets ‘Wine quality_red’ and ‘Wine quality_white’ are included in the same database ‘Wine Quality Data Set’. Limited to data quality, ‘processed.cleveland’ and ‘poker-hand-training-true’ two subsets were selected as represents of the databases ‘Heart Disease Data Set’ and ‘Poker hand data set’, respectively

ISSN: 1475-925X