OPTIMASI PARAMETER LOGISTIC REGRESSION UNTUK KLASIFIKASI WEB BERBAHAYA MENGGUNAKAN ALGORITMA FIREFLY

Abstract

The development of technology and information, especially the internet in Indonesia, has experienced rapid growth that has changed the pattern of information consumption among the public. Websites have now become the main means for communication, transactions, and learning. However, the increased use of the web also poses security risks in the form of malicious websites that are often exploited by cybercriminals to steal data or spread malware. Detecting malicious websites becomes a challenge because their appearance resembles normal sites. This research proposes Optimization of Logistic Regression Algorithm Parameters using the Firefly Algorithm to improve accuracy in classifying malicious websites. The dataset used comes from Kaggle with a total of 1,781 data points and 21 attributes. The research process includes data cleaning, normalization using Min-Max Scaler. Data is divided into training and testing sets with several data splits (60:40, 65:35, 70:30, 80:20, 90:10), and evaluation using a confusion matrix. The test results show that the application of the Firefly algorithm can improve the accuracy of the Logistic Regression model, with Firefly enhancing the model's performance in detecting malicious websites.