CNN-Based Topic Modeling and Semantic Sentiment Analysis Using Machine Learning

Big data's complexity and heterogeneity provide substantial challenges for businesses. Large volumes of data provide a basic challenge for organizations. Methods for processing massive amounts of unprocessed information in different forms also need to be improved. Making data storage that can grow as needed to store and retrieve vital information is also important. We proposed a new approach which utilizes deep learning (DL) and machine learning (ML) to analyze sentiment datasets. In this research, a CNN was used to do initial topic modeling using a sentiment dataset that had been standardized using POS tagging with NLTK. Random forest (RF) was used with Logistic regression (LR) to choose the best features. Ensemble Classifier utilizes many machine learning techniques, including decision tree (DT), support vector machine (SVM), logistic regression (LR), and naive Bayes (NB), to determine which model is most accurate at predicting the sentiment category. Maximum Sentiment classification prediction accuracy using proposed ensemble classifier is about 95%. Similarly, precision and recall also achieved better than existing works (i.e.) up to 95%.