This repository contains the final project for the CSE523 Machine Learning Course.
With the advancement of social media platforms and its growth there is a huge volume of data present in these platform. Internet has brought the world so close together and provided people a means to express themselves. Social Networking sites like Twitter, Facebook, Instagram are gaining popularity among people as they allow users to express their opinions on variety of topics, have discussions and post messages. This gives rise to opportunities to detect and predict sentiments of people on various topics. There has been a lot of work in the field of sentiment analysis of twitter data. This project mainly focuses on sentiment analysis of twitter data, which is helpful to analyze the information in the tweets where the opinions are highly unstructured and heterogeneous. The aim of this project is to come up various machine learning models to accurately detect sentiment of a tweet.
As a part of CSE523 Project, we've built a highly robust classifier, capable of classifying a tweet into positive or negative sentiment.
Classifier algorithms used:
- Multinomial Naive Bayes Algorithm
- Logistic Regression Algorithm
Vectorizer algorithms used:
- TF-IDF Vectorizer
- Count Vectorizer
This section contains the result of the Naive Bayesian Classifier (implemented from scratch)
- http://www.ijstr.org/final-print/may2020/A-Literature-Survey-On-Sentiment-Analysis-Techniques-Involving-Social-Media-And-Online-Platforms.pdf
- https://cofounderstown.com/twitter-sentimental-analysis-using-naive-bayes-9eb73
- https://arxiv.org/ftp/arxiv/papers/1607/1607.07384.pdf
- https://www.researchgate.net/publication/327251557\_Depression\_detection\\\_from\_social\_network\_data\_using\_machine\_learning\_techniques
- https://web.stanford.edu/~jurafsky/slp3/5.pdf