A machine learning system that takes a comment and classifies it as offensive or non-offensive (neutral). This system will be trained in a data set with comments in which the tags (insult or non-insult) are known. Classification algorithms used: Naive Bayes, SVM, Random Forest.

Fake news detection

Built using Python, Streamlit, and NLTK, the Hate Speech Detection App employs a Decision Tree Classifier for identifying hate speech in text. It features real-time speech input, NLP preprocessing, and a user-friendly Streamlit interface, offering both visual and text-to-speech result presentation.

This work proposes a new approach for detecting fake news in datasets using some of the most widely used algorithms of machine learning.

Market trends and investment insights

Sentiment Analysis is the process of determining whether a piece of writing is positive, negative or neutral.

💉 Vaccine Sentiment Classifier is a deep learning classifier trained on real world twitter data, that distinguishes 3 types of tweets: Neutral, Anti-vax & Pro-vax.

The movie recommendation system is implemented using content based filtering

A spam email chacking system using the Complement-Naive-Bayes algorithm on SpamAssassin datasets

A Google Chrome Extension that estimates the Reliability, Polarity and Subjectivity of any news article on the web. It allows you to like/dislike any article and recommends you articles based on your choices.

Using Multinomial Naive Bayes Classifier to classify SMS messages as SPAM or HAM. Techniques used include Count Vectorizer and Text mining using TF-IDF.

Content-based recommendation engine using Python and Scikitlearn, using concepts of Cosine distance and Euclidean distance. Finally, by using IMDB 5000 movie dataset built a content-based recommendation engine using CountVectorize and Cosine similarity scores between movies.

AI-powered classifier mobile app using NLP to spot fake job ads and protect users from online scams. Our system analyzes language patterns and leverages algorithms to create a safe and trustworthy job search experience.

The scope of this project is to classify fake and true news. After performing an analysis on the dataset using two different vectorizers and two machine learning algorithms, the results are conveyed in the form of accuracy score and confusion matrices.

This repository houses 3 different Jupyter Notebooks that each analyze the similarity in data points to most effectively inform customer recommendations in the retail space.

Classifying a tweet as positive, neutral, or negative sentiment using Natural Language Processing (CBOW approaches) and Traditional Machine Learning Algorithms.

Case Study- Text Mining

Case Study- Text MIning