<![CDATA[The rapid digital transformation of Indonesian MSMEs has increased exposure to cyber threats such as phishing, malware, ransomware, and DDoS. This study aims to review the application of data mining and machine learning for cyber threat detection, especially in the context of small businesses. A systematic literature review (SLR) with the PRISMA approach was used. Articles were collected from Google Scholar and Semantic Scholar (2018–2026), screened using inclusion and exclusion criteria, and analyzed descriptively. From 19 collected articles, 11 were reviewed: seven on data mining/machine learning for cyber threat detection and four on MSME cybersecurity. The results show that Random Forest, C4.5/J48, SVM/SMO, Naive Bayes, and K-Means are commonly used. Feature selection (Information Gain, PCA) and class balancing (SMOTE) improve accuracy and training efficiency, with reported accuracy generally above 90%. However, most studies use public datasets such as KDD Cup’99, AWID, CICIDS2017, and CIC-MalMem-2022, not data from MSME environments. Key challenges include class imbalance, false positives, limited resources, data quality, scalability, and low digital literacy. Non-technical factors—management commitment, governance, training, and cost—are also critical. Data mining has potential for MSME threat detection, but implementation requires lightweight, context-specific models and organizational readiness.]]>
Copyrights © 2026
