ENHANCING SOFTWARE MIGRATION THROUGH MACHINE LEARNING (ML) AND ARTIFICIAL INTELLIGENCE (AI
Keywords:
Software Migration, Machine Learning (ML), Artificial Intelligence (AI), Cloud Migration, Data Transformation, Anomaly Detection, Code Optimization, Migration AutomationAbstract
This study explores the application of the Random Forest classifier for analyzing and predicting software
configuration patterns, particularly in API migrations. Random Forest is an ensemble learning method that builds
multiple decision trees on various randomly selected subsets of the dataset and aggregates their outputs to form a
final prediction. This approach significantly enhances predictive performance by reducing over fitting and
increasing generalization. In this work, Random Forest is applied to a dataset of Java API migrations, where features
such as migration date, code structure, and transformation details are used to predict confidence scores and classify
the success of configuration changes. The results demonstrate that Random Forest provides high accuracy and
robustness in predicting software migration outcomes. Its voting mechanism, based on averaging the outputs of
multiple trees, leads to reliable and consistent predictions. This suggests that Random Forest can be a valuable tool
in automating and improving decision-making in software configuration management tasks.
