Publication Search

This study aims to analyze and compare the performance of the Naive Bayes, K-Nearest Neighbors (KNN), and Decision Tree algorithms in predicting the purchase intention of e-commerce visitors using the Online Shoppers Purchasing Intention Dataset, which consists of 12,330 records and 18 variables, with the Revenue variable serving as the classification target. The preprocessing stage involved transforming categorical and boolean variables into numerical form, standardizing features using StandardScaler, and splitting the dataset into 80 percent training data and 20 percent testing data. Model evaluation was conducted using accuracy, precision, recall, F1-score, and ROC-AUC metrics, and was further strengthened by 10-fold cross-validation to obtain more stable results. The findings indicate that KNN achieved the highest accuracy of 0.866180, while Naive Bayes produced the highest recall value of 0.690998 and the highest ROC-AUC value of 0.821696. Meanwhile, Decision Tree demonstrated relatively balanced performance with an accuracy of 0.857259 and an F1-score of 0.571776, whereas the cross-validation results identified KNN as the model with the highest average accuracy of 0.8770. These findings suggest that the selection of a classification model for purchase intention prediction cannot rely solely on a single evaluation metric, as each algorithm possesses different strengths. Therefore, a comparative approach among algorithms can help determine the most suitable model for supporting consumer behavior analysis on e-commerce platforms.

This study analyzes the prediction of poverty levels in North Sumatra Province by applying the Long Short-Term Memory (LSTM) method based on time series integrated with Google Earth Engine (GEE). Historical poverty data of districts/cities were obtained from the Central Statistics Agency (BPS) and processed using Python in Google Colab for LSTM model training. The prediction results are visualized spatially in the form of thematic maps through GEE to identify areas with high poverty rates. The evaluation model was carried out by calculating MAE, RMSE, MAPE, and prediction accuracy, with most areas having an accuracy above 80%. These findings indicate that this approach is effective in mapping poverty trends and supporting data-driven policies. This predictive model can be the basis for more targeted social interventions and strategies for developing inclusive and sustainable regional development.

The bodywork industry in Indonesia is under high competitive pressure, requiring companies to be more adaptive in understanding customer behavior in order to maintain business continuity. PT. Bengawan Karya Sakti as one of the national bodywork companies, has not optimally utilized historical transaction data to assess customer loyalty. This study aims to identify customer loyalty segmentation through the application of the RFM (Recency, Frequency, Monetary) method, which is used to analyze sales transaction data in 2022 and 2023. The study uses the CRISP-DM approach which includes the stages of business understanding, data exploration, data cleaning and processing, modeling, evaluation, and implementation of results. The transaction data analyzed includes attributes of transaction date, customer, number of transactions, and transaction value, which are then processed into RFM scores based on the transaction year and classified into categories such as Very Loyal, Loyal, At Risk, and others. The segmentation results show an increase in the number of very loyal customers from 2022 to 2023, as well as a significant decrease in inactive and at-risk customers. The chi-square statistical test shows that the difference in customer distribution between years is statistically significant (p-value <0.05), indicating a real influence from the company's strategy or external factors. The main conclusion of this study is that the RFM method is effective in the bodywork industry to support data-based marketing decision making and more targeted customer retention strategies.