<![CDATA[ABSTRAKPertumbuhan Internet of Things (IoT) meningkatkan konektivitas perangkat sekaligus memperluas risiko serangan botnet. Keterbatasan sumber daya pada perangkat IoT menyebabkan metode deteksi tradisional kurang efektif dalam mengenali pola serangan yang dinamis. Penelitian ini mengusulkan model deteksi botnet berbasis Long Short-Term Memory (LSTM) dengan memanfaatkan karakteristik temporal trafik jaringan. Dataset yang digunakan merupakan subset Bot-IoT sebanyak 14.000 data yang terdiri atas trafik normal dan serangan botnet. Tahap pengolahan data meliputi pembersihan data, seleksi fitur, normalisasi Min-Max, dan pembentukan sequence. Hasil pengujian menunjukkan bahwa model LSTM mencapai akurasi 95,89% dan nilai AUC 0,97. Temuan ini menunjukkan bahwa LSTM efektif untuk mendeteksi aktivitas botnet pada lingkungan IoT dengan keterbatasan sumber daya komputasi.Kata kunci: Internet of Things, Botnet Detection, Deep Learning, Long Short-Term Memory, Intrusion Detection SystemABSTRACTThe rapid growth of the Internet of Things (IoT) has increased device connectivity while expanding the risk of botnet attacks. Limited computational resources in IoT devices reduce the effectiveness of traditional detection methods in identifying dynamic attack patterns. This study proposes a Long Short-Term Memory (LSTM)-based botnet detection model by leveraging the temporal characteristics of network traffic. A subset of the Bot-IoT dataset consisting of 14,000 normal and botnet traffic records was used. Data preprocessing included cleaning, feature selection, Min-Max normalization, and sequence construction. Experimental results show that the proposed LSTM model achieved an accuracy of 95.89% and an AUC of 0.97. These findings indicate that LSTM is effective for detecting botnet activities in IoT environments with limited computational resources.Keywords: Internet of Things, Botnet Detection, Deep Learning, Long Short-Term Memory, Intrusion Detection System]]>
