Name: Sowmya Vijeth
Type: User
Company: TheMavericksIndia
Bio: Interested and enthusiastic about predictive data analytics using pandas, scikit learn, ML and DL techniques . Currently pursuing masters from LJMU
Location: Bengaluru
Sowmya Vijeth's Projects
A WEKA package for analyzing emotion and sentiment of tweets.
This repos contains notebooks for the Advanced Solutions Lab: ML Immersion
Web Scraping and Analyzing Performance of Your Content on LinkedIn.
A real-time interactive web app based on data pipelines using streaming Twitter data, automated sentiment analysis, and MySQL&PostgreSQL database (Deployed on Heroku)
Online reviews play a pivotal role in helping people purchase products eventually influencing the business verdicts. Online reviews have led to fake review writing, which can either be paid human writers or machine generated deceptive reviews with the aim to influence future customers opinion. In this project we have tried to tackle this problem with the help of a classifier that processes the review text and userβs behavioral pattern as input and predicts whether the review is genuine or not. The learning algorithms we experimented include logistic regression, multinomial Naive Bayes, K-Nearest Neighbour (KNN), Random Forest classifier, Convolutional Neural Network (CNN) and CNN with Long short term memory (LSTM). From the results of our experiment we can see that Logistic Regression and KNN performed better with approximately 60-64% accuracy and Naive Bayes and Random Forest Classifier (RFC) does not work well with our dataset with only 50% accuracy and CNN-LSTM does not give a very high accuracy (21%) i.e. but has a good recall of 0.82.