NewsClassifier_And_Online_news_fetcher

News Classifier aims to classify news depending on the several categories of news present around the world.

---

Complete Walkthrough of the process is documented in my Medium Article, Do have a look at it from this link.

The app has been containerized using Docker and deployed on AWS using Fargate.

Data

News data obtained from Kaggle. link

Objective

• Objective of this project is to help apps automatically classify news.
• People can see news depending on their choice/Interests.
• Preventing fake news from spread by predicting the fake news and not letting it go to the customers.

---

Dependencies

• re
• sys
• pandas
• numpy
• matplotlib
• seaborn
• nltk
• contractions
• autocorrect
• sklearn
• zeugma

---

Models

• Logistic Regression
• MultinomialNB
• SVM

---

Metrics

• accuracy_score
• matthews_corrcoef

---

Performance

• Logistic Regression
  • accuracy_score -----> 0.65
  • matthews_corrcoef --> 0.63
• MultinomialNB
  • accuracy_score -----> 0.44
  • matthews_corrcoef --> 0.39

---

Result

1. Predict Only
   

2. Predict and fetch news online
   

3. Manual Search online