For the new_headline_sentiment.csv, there are sentiment scores with headlines. But I don't understand how to get this sentiment score if I want to expand this data size for this file.
Hi,
In the initial section, we have used a dataset that already contains the sentiment score as there are data providers who provide sentiment class and scores. If you want to calculate the sentiment score, you can try two approaches.
- ML-based
- Lexical based which works on dictionaries of scored lexicons.
Further in the course, we have mentioned how word embedding approaches work. And finally, you can use BERT natural language processing model and XGBOOST to create a trading strategy based on sentiments.
Hope this helps.