Klasifikasi Data Mining Pengguna Instagram Terhadap Penurunan Jumlah Followers Akibat Penghapusan Bot & Akun Palsu Menggunakan Algoritma Naïve Bayes

Abstract

The development of social media, particularly Instagram, has made the number of followers an indicator of popularity, credibility, and the effectiveness of information dissemination for both individuals and businesses. However, Instagram's policy of periodically removing bots and fake accounts has caused some users to experience a decrease in their follower count, which impacts engagement and account performance. This study aims to apply data mining classification techniques using the Naïve Bayes algorithm to classify Instagram users based on the causes of follower decline due to bot and fake account removal. The research data was obtained from the collection of Instagram user reviews on the Google Play Store and App Store, which underwent data selection, data cleaning, data transformation, and sentiment labeling into positive, negative, and neutral categories. The data processing was carried out using RapidMiner Studio software, with preprocessing, dividing training and test data, building a Naïve Bayes classification model, and evaluating using a Confusion Matrix. The results of this study indicate that the Naïve Bayes algorithm is capable of classifying data with a high level of accuracy in identifying user sentiment related to the decline in followers due to the removal of bots and fake accounts. The resulting model provides useful information regarding user perceptions of Instagram's policies, which can be used as evaluation material to improve the quality of social media services. Thus, the application of the Naïve Bayes algorithm has proven effective in classifying Instagram user review data and can serve as a reference for further research using other classification algorithms to achieve more optimal performance.