PEMODELAN KLASIFIKASI PREFERENSI IDEOLOGI PARTAI POLITIK PADA PEMILIH PEMULA DI PROVINSI BANTEN

Abstract

First-time voters in the 2024 election are expected to play an important role in determining the future direction of Indonesia, by bringing fresh perspectives and new aspirations for change amidst increasingly complex political dynamics. This study aims to apply classification modeling using political party preference data on first-time voters in Banten Province. Data were collected through a survey with responses from 3 different party ideologies, namely the Secular National Party (PNS), the Religious National Party (PNR), and the Islamic Party (PI). The analysis stages include data preprocessing, forming a Random Forest classification model, evaluating model performance using accuracy, precision, recall, and F1 score measures, and mapping the importance of features from the Random Forest results. The SMOTE technique is used to handle class imbalance, where the majority of party ideologies from the data obtained are PNS, which is 50%. The results obtained from 10-fold cross-validation with multiclass classification show an accuracy of 64.00%, with precision, recall, and F1 score values of 61.88%, 62.33%, and 60.41%. The variable importance of political party ideology is the ideological background of the voters themselves, with a Mean Decrease Gini (MDG) score is 91.28.