Sentimental Analysis of Movie Review Based on Naive Bayes and Random Forest Technique

Rapid advancements in text classification algorithms have kept pace with the exponential rise of digital materials in recent years. In order to automatically extract expressive features, new machine learning algorithms have been suggested that take advantage of recent developments in deep learning techniques. Rapid progress in this area has spawned a multitude of ways for translating human speech into machine-readable data. Ad hoc pre-processing processes are utilized in conjunction with state-of-the-art language modelling algorithms; nevertheless, their presentation is generally glossed over in favor of a more thorough explanation of the classification stage. This work aims at building a model that is used to analyze sentiments, where accuracy is tested by taking data sets of positive and negative movie reviews. Proposed work is divided into 3 tasks: Data Extraction, Preprocessing and Modelling. We have used text classification techniques of Machine Learning by making changes in text vectorization using BOW (Bag Of Words), N-grams and TFIDF and modelling is built using Naïve Bayes and Random Forest algorithms. Our findings demonstrate that the suggested model provides superior performance in terms of accuracy, precision, recall, and f-measure. Using cutting-edge methods, proposed model achieved competitive results on the IMDB movie review dataset.