This document presents a study on automatic text classification of news blogs using machine learning. The study uses a training dataset of 20 lakh news articles to train a machine learning model. It explores techniques like preprocessing, vectorization using count vectorizer, normalization using TF-IDF, and classification using stochastic gradient descent algorithm. The SGD algorithm achieved the highest accuracy on test data compared to other algorithms like naive Bayes and logistic regression. The trained model was able to accurately classify new news articles into different categories.