Overview / Usage

We used Headlines, Descriptions, and a combination of descriptions and headlines and try to predict the Sentiment of the News (i.e. Positive, Negative, Neutral )

This project is profitable for the company as:

1. The company can keep a check on the nature of the content they are publishing to balance out the mood of the reader.
2. It can also help them structure the articles in proper order.
3. We can also get a rough idea of financial condition with the Financial News record sentiment.
4. This can also help in the filter if the user wants to see particular sentiment news.

Methodology / Approach

I used Supervised Machine Learning along With NLP.

1. Cleaning the Data using basic **EDA **null values handling and missing values handling.
2. Then performed preprocessing of **NLP **.
3. Performed Bag-of-words and Tf-Idf for NLP.
4. Used multiple Classifiers to choose the best among them.
5. A classification Report is used and Accuracy is chosen for the best model selection.

Technologies Used

Libraries:

• NLTK
• Sklearn

Repository

https://github.com/deepakjoshi2k/Financial_News_Sentiment_Analysis

Other links

• Github Profile (https://github.com/deepakjoshi2k)