Scholarlink

<p>According to research, Android-based malware has increased at an alarming rate and is capable of hiding in the Android system using various obfuscation techniques. Detection methods such as intrusion detection systems (IDS) are rapid and efficient in malware detection. In this paper, a hybrid IDS based on artificial intelligence methods is presented to detect malware. These IDSs consist of two steps: (1) feature selection and (2) the detection of malware. In the feature selection stage, the data set is pre-processed, and then the feature selection is made using the GWOPSO combination method. In this method, the search is performed using the gray wolf algorithm (GWO), and features are selected using the Particle Swarm Optimization (PSO) algorithm. This balances between exploration and exploitation as well as speed up the execution. Secondly, malware is detected using GA_Fuzzy_K-Modes. In this method, the K-Modes algorithm is used to detect malware, increases convergence speed, and diversify the population before the crossover process, and the GA algorithm is used to achieve the global optimal solution. This IDS was evaluated using two Drebin and CICMalDroid datasets, and the results of the evaluation have shown that the proposed method was able to identify with an accuracy of 0.994 and 0.981.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

GAFuzzyK-modes method and hybrid metaheuristics approach for feature selection and android malware detection

  • 期刊论文

  • 发表日期：

    2026年7月14日

查看全文
  • Mahdieh Maazalahi,
  • Soodeh Hosseini

摘要

According to research, Android-based malware has increased at an alarming rate and is capable of hiding in the Android system using various obfuscation techniques. Detection methods such as intrusion detection systems (IDS) are rapid and efficient in malware detection. In this paper, a hybrid IDS based on artificial intelligence methods is presented to detect malware. These IDSs consist of two steps: (1) feature selection and (2) the detection of malware. In the feature selection stage, the data set is pre-processed, and then the feature selection is made using the GWOPSO combination method. In this method, the search is performed using the gray wolf algorithm (GWO), and features are selected using the Particle Swarm Optimization (PSO) algorithm. This balances between exploration and exploitation as well as speed up the execution. Secondly, malware is detected using GA_Fuzzy_K-Modes. In this method, the K-Modes algorithm is used to detect malware, increases convergence speed, and diversify the population before the crossover process, and the GA algorithm is used to achieve the global optimal solution. This IDS was evaluated using two Drebin and CICMalDroid datasets, and the results of the evaluation have shown that the proposed method was able to identify with an accuracy of 0.994 and 0.981.