Supervised Code-Mixed Data Sentiment Analysis

Sentiment Analysis of data extracted from social media is a very important and necessary task gaining popularity for many years applicable on various fields like opinion mining. Analyzing code-mixed English data is a typical task due to the complication of dataset. Machine Learning procedures has paved the way to handle these datasets with utmost ease and efficiency. In this research paper Sentiment Analysis on code-mixed tweets of English by means of different Machine Learning Classification techniques is implemented after pre-processing the gathered data using Count Vectorizer and Tf-Idf Vectorizer. The classification of a tweet or tweets into subjectivity, polarity and its analysis is our main objective. Supervised learning models like SVM, Decision Tree, Logistic Regression and K-nearest neighbor are instigated for analyzing the tweets and the result based on evaluation metrics like F1-score, and accuracy is used to find the most appropriate and efficient technique.