Model Machine Learning Untuk Klasifikasi Kemiripan Komik Menggunakan Jaccard Similarity Dan Logistic Regression

Abstract

TeenToon is a mobile-based digital comic platform that provides digital comics that incorporate Christian values. As the number of available comics continues to grow, the platform faces the challenge of information overload, making it difficult for users to discover comics that match their preferences. This study aimed to develop a comic recommendation feature for TeenToon using a Content-Based Filtering approach by applying Jaccard Similarity and Logistic Regression. The similarity between comics was calculated based on genre and tag attributes, and the recommendations were presented as the Top 5 recommendations. The study utilized 25 comic titles, which were paired to generate 300 data pairs. Each pair was labeled by three validators based on subjective assessments of comic similarity, considering both the storyline and the similarity of genre and tag attributes. The final labels were determined using majority voting. The Jaccard Similarity score was then calculated and used as an input feature for the Logistic Regression classification model, which was subsequently implemented on the comic detail page of the TeenToon application. The results showed that the Logistic Regression model achieved an accuracy of 88.46% and a precision of 71.43%. These findings indicate that Logistic Regression is effective for developing an attribute-similarity-based recommendation system for the TeenToon application.