PQTD-TM: Integrating Information Retrieval with Machine Learning for effective Placement Topic Discovery

Technical interviews, particularly coding rounds, are pivotal in shaping career trajectories, especially for roles in top-tier companies like Amazon, Google, and Microsoft. Understanding the patterns and nuances of coding problems and their associated tags is essential for effective interview preparation. To aid candidates in excelling in these assessments, we analyzed coding problems encountered in such interviews. We compile a comprehensive dataset of coding problems and their associated tags, sourced from real technical interviews, ensuring reliability and relevance. We propose a topic prediction approach for placement related questions, PQTD-TM (Placement question topic discovery using TF IDF and Machine Learning). The proposed technique integrates information retrieval technique (TF-IDF) with six machine learning classifiers for prediction. We evaluated the predictive performance of various machine learning models, including Logistic Regression, Decision Trees, Naive Bayes, KNN, One-vs-Rest Classifier, and Random Forest. Random forest attained the best prediction accuracy of 93%