Opinion Mining with Hypertuned Random Forest

Opinion Mining with Hypertuned Random Forest

Arpita Ashok Khot, Dhruvin Hasmukh Barot, Praveen Raju Bhandari

Abstract - E-tailing has seen a lot of development, while brick-and-mortar had a significant downfall. A few reasons for that was that lack of traveling, comparisons between best prices, offers, sales etc. A beneficial feature in online shopping is the reviews about the product. The purchasers give their product experience which helps other consumers to select what is best suited for themselves. But the purchasers are in huge amounts and thus their reviews are also in greater quantity, therefore it becomes very difficult for the consumer to find the Positive, Neutral and Negative reviews. Here Machine Learning comes into picture, by using various ML models one can segregate all types of reviews and classify them into different sections so that the consumer can directly get a comparative analysis about his/her product. This is known as Opinion Mining also commonly known as sentimental analysis. Applications are that its uses cut across sectors, aiding market research, customer satisfaction enhancement, and decision-making. Businesses and organizations may gain a competitive edge and better serve the requirements of their target audience by comprehending and utilizing the power of public opinion. The Fundamental idea of this paper is to discuss the sentimental evaluation and categorize them into their respective polarities based on the given sentence.

Keywords: Sentimental analysis, Positive, Neutral, Negative, Natural Language Processing, Random Forest, SVC, Logistic Regression, Gaussian Naive Bayes, KNN, Grid search view, Random Search, Cross Validation, Lemmatization,Tokenization, Vectorization  .