Hi, I bought course Nataural language Processing, and in thier examples (.csv) these have included the sentiment_class and sentiment_score. I understand how calculate the sentiment_class and sentiment _scores, but my question is, I fetched articles from Webhose.io API, to put the sentiment_class this is manual (how I do it). thanks for you help.
Let me get this right. You have been able to use the provided dataset with sentiment score and class. You have downloaded similar data from webhose and now intend to get a sentiment score for the same. Do correct me if I am wrong.
To calculate the sentiment score, there are two methods:
1. ML-based but they need already labelled data.
2. Lexical based which work on dictionaries of scored lexicons.
Since we do not already have labelled data we will follow the Lexical approach. Vader is a popular library which is used to calculate sentiment score using the lexical method. This link has examples and documentation on Vader. I would recommend this method to calculate the sentiment score because it does not need explicit labelling. However, the dictionary in vader is made from twitter text so it does not always work well with financial news data. One way to go around this is to extend Vader's dictionary with financial terms.
Another Quantra course which is a part of the learning track Sentiment Analysis in Trading extensively covers this library to calculate sentiment score and use it in intraday trading.
Another Quantra course which is a part of the learning track Sentiment Analysis in Trading extensively covers this library to calculate sentiment score and use it in intraday trading.
Thanks for you help