Sentiment Analysis Using News

News articles are a great source of information to gain data about the current market trends. But it is difficult to gain data from the news. Here’s a way to fetch data from the news - “Sentiment Analysis”. This video discusses the basics of Sentiment Analysis and how to fetch accurate data from news using sentiment analysis. There are several challenges when you try to generate a sentiment score for news articles just like we did for tweets. The problem is there are many words so it may not give you a proper score because there are many many words that are talking about apple and its employees. If it is a tweet it directly talks about apple but if it is an article it might talk about apple employees' apple performance 10 years ago. Your algorithm must be smart enough to understand what is the context in it and how many like it shouldn't give you bad results just because the article is big so that's the reason in this course we have used only headlines. We have used headlines and tried to generate sentiment scores but to improve this you can see in another course named “Natural language processing and trading”. This is an ex-natural extension of this particular course where we try to design a machine learning model that learns the sentiment score given by this better model and tries to analyze the articles. We have used the XG Boost model a machinery model to learn the logic with which we use the better model so this is the most basic form of using machine learning to underst