A New Sentiment Analysis System Of Tweets Based On Machine Learning Approach

A very huge amount of data is generated every second for microblogs, content sharing via Social media sites and social networking. Twitter is an important popular microblog where people voice their opinions with regard to daily issues. Recently, analyzing these opinions is the main concern of Sentiment analysis (or opinion mining). Efficiently capturing, gathering and analyzing sentiments has been challenging for researchers. To deal with these challenges, in this paper we propose a highly accurate model for sentiment analysis of tweets. Using the Crowdflower's dataset, we started by data preprocessing (replace missing value, Denoising, tokenization, stemming...). We applied a semantic model with Term Frequency, Inverse Document Frequency weighting for data representation. In the measuring and evaluation step we applied four machine-learning algorithms such as Naive Bayesian, K-Nearest Neighbors, Neural Networks (LSTM), and Support Vector Machine. Afterwards, and based on the results we boiled a highly efficient prediction model with python, we trained and evaluated the classification model according to the most efficient metrics measures in this field, then tested the model on a set of unclassified tweets, to predict the sentiment class of each tweets. Experimental results demonstrate that our model reached a high accuracy compared to the other models.