Enrich Skills Recommendation Based on Sentiment Analysis Using Ensemble Learning
Muthu S. Nidhya, Vijaya Kumar Guivada, Subramanian Pitchiah Maniraj, Neetu Pillai · 2024
Recent research fields that are closely related to one another include sentiment analysis, emotion recognition, and text data emotion detection. Sentiment analysis looks for and attempts to discover neutral, favorable, or unfavorable opinions in text. In this study, we propose a machine learning (ML) technique advice on enrich IT skills that an employee needs to be learning based on his competency or ability or background of information technology. Sentiment analysis, reviewer clustering, semantic network grouping of related sentences, and recommendation system are among the steps. With the use of parts of speech tags, the semantic network joins similar lines at the first level using parts of speech tags and other sources. Using the “bag of words” and term frequency-inverse document frequency feature extraction techniques, keywords are extracted from the pre-processed data. Training and testing phases of sentiment analysis are carried out at the second level using ensemble learning. The outcomes of this stage are passed on to the clustering process, which divides the population into groups according to age, geography, and gender. Using accuracy, precision, recall, F-1 measure, the model is assessed to the highest standard. Comparing the enrich skills recommendation system to the existing algorithms of SVM, CNN, ANN, LSTM, and Bi-directional (BI)-LSTM, our new suggested hybrid approach of butterfly optimization with Naive Bayes algorithm is intended to deliver the best level of overall efficiency.