Open AccessAttribute analysis with classification algorithm on election participation
Author(s) -
Arif Senja Fitrani,
Mochamad Alfan Rosid,
Fajar Muharram,
F L Kodriyah
Publication year - 2020
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/821/1/012034
Subject(s) - naive bayes classifier , class (philosophy) , rank (graph theory) , government (linguistics) , commission , voting , computer science , political science , presidential system , mathematics , artificial intelligence , politics , support vector machine , linguistics , philosophy , combinatorics , law
Stages of General Elections of the President, DPD, DPR, Provincial DPRD and Regency / City DPRD in Indonesia are determined by institutions namely the General Election Commission (KPU), where there is a measure of success in holding direct, general, and free. Another component of the implementation of elections is that there are contestants and voters. In the voter factor, this is also a measure for success in the overall process of implementation, namely success if high community participation in the administration of elections. However, vice versa, if community participation is low, one of them is the level of public confidence in the organizers (government) decreases. Data mining classification analysis and modification of attributes in prediction classes “Hadir” and “Tidak Hadir” on the final voter list (DPT). The number of datasets is 4249 instances, and the number of attributes is 11. The percentage results are 89.3417% for the Naive Bayes algorithm for prediction classes in the Presidential Election, DPD, DPR, Provincial DPRD and Regency/City DPRD in 2019. Further analysis is done on eliminating some attributes to obtain information, whether it has a significant effect on the results of predictions. And in this analysis, for the ten attributes with the removal of “statusKK” (highest rank gain ratio) for the prediction class, the results are the worst. = After nine attributes, the removal of the “rt” and “tps” attribute (second and third rank) for the prediction class is the best result. There is the highest percentage difference for the prediction class on the classification algorithm for modified and or unmodified status attributes, the results of the percentages and classification algorithms are different.