Fake Review Detection in Online Customer Reviews Using Unsupervised Machine Learning Techniques

Detection of Fake Reviews on Flipkart using Machine Learning Architectures.

The dataset used includes 15166 different reviews uploaded on the flipkart website.

The following steps are performed:

1. Pre-processing of dataset
2. Sentiment Analysis on Review Title and Review Text
3. Customer Segmentation and performing K Means Clustering
4. EDA on clusters formed
5. Anomaly Detection Using Isolation Forest
6. Unsupervised learning algorithms to divide the review data into clusters and perform detailed analysis of the clusters to categorize reviews as fake or real.