<![CDATA[The development of digital platforms in the Micro, Small, and Medium Enterprises (MSMEs) sector generates various types of user activity data that have the potential to support data-driven decision-making. However, most MSME digital platforms still primarily focus on marketing and transaction functions, while the utilization of consumer behavior data for predictive analysis remains limited. This study aims to implement machine learning to predict consumer behavior on the khas-cilacap.com digital platform based on customer analytics. This study employs a quantitative approach using a computational experimental method. The dataset used is dataset consisting of 350 user records designed based on digital platform user activity characteristics, including visit frequency, product interaction, session duration, shopping cart activity, transaction frequency, transaction value, and customer status. The research stages include data preprocessing, dataset splitting into training and testing sets, classification model development using Logistic Regression, Decision Tree, and Random Forestalgorithms, and model evaluation using accuracy, precision, recall, and F1-score metrics. The results show that the Random Forestalgorithm achieved the best performance compared to other algorithms, with an accuracy of 91.25%, precision of 90.80%, recall of 89.70%, and F1-score of 90.20%. Feature importance analysis indicates that Transaction Frequency and transaction value are the most influential factors in identifying potential customers. These findings demonstrate that the implementation of machine learning-based Customer analytics can support the development of khas-cilacap.com as a digital MSME platform capable of generating predictive insights for marketing strategies and data-driven decision-making]]>
Copyrights © 2025
