Quality Prediction of a Stack Overflow Question Using Machine Learning

The support forums for developers are becoming more and more popular. Despite the fact that many programmers view adaptable knowledge as a precious asset, there could be problems with the accuracy of the information obtained. Morphological segmentation processes have progressed quickly as a consequence of the massive increase in digital resources. It may be difficult to distinguish between the quality of the questions and all of their substance on Q&A platforms. One such prevalent Q&A website that is important for programmers, developers, researchers, etc., to get technical information, especially in the field of Computer Science is Stack Overflow. Recently emerging machine learning optimization techniques, which make use of the most significant advancements in sophisticated learning methodologies, enable the autonomous retrieval of expressive characteristics. There are multiple methods for transforming human language into system-interpretable data as a result of the technological advancement in these approaches. This research presents various techniques that are implemented to predict the quality of questions. To analyze the textual data of 60,000 questions, the Lexicon-based Sentiment Analysis concept of Natural Language Processing (NLP) is utilized. Optimized methods like Naive Bayes (NB), Support Vector Classifier (SVC), K-Nearest Neighbors (KNNs), Decision Tree (DT), Random Forest (RF), and Logistic Regression (LR) are employed on the model. Finally, a relative and absolute analysis is performed with a major emphasis on accuracy and novel results from various approaches.