<![CDATA[This study is motivated by the limited utilization of AI-based metrics to predict task success among developers in software development projects. The main issue addressed is the absence of a systematic comparative approach to classification algorithms in identifying the most effective model in this context. Therefore, this research compares the performance of three classification algorithms—K-Nearest Neighbors (KNN), Decision Tree (DT), and Naïve Bayes (NB)—in predicting task success using AI-metrics data. The evaluation metrics include precision, recall, F1-score, and accuracy, presented through classification reports and confusion matrices. The results show that DT achieved an accuracy of 91%, KNN 92%, and NB 86%. The confusion matrix analysis indicates that DT demonstrates high precision, KNN shows minor imbalance, and NB struggles to identify minority classes. Additionally, clustering was performed using the K-Means algorithm and visualized in two dimensions through Principal Component Analysis (PCA), revealing clear segmentation among developer groups. The ultimate benefit of this study is to provide a foundation for decision-making in selecting the most appropriate algorithm to enhance developer team effectiveness and personalize managerial strategies. The novelty of this research lies in the combined application of classification and clustering approaches using AI-metrics to more accurately and datadrivenly identify developer task success. ]]>
Copyrights © 2025
