0
$\begingroup$

I'm trying to build a model which gives me the sentiments of the Financial News related to a company and I want to predict the stock price accordingly. But the major problem that I'm facing is understanding the news for the counterpart.

Let's say I have a news headline as "Total is under pressure and CEO has confirmed that they are planning to sell stakes soon". My model will always give negative sentiment which is correct, but this might actually be a good news for Shell or Exxon, lets say. But how do I tell my model that it is actually a good news for Shell.

Also is there any good process to understand which news relate to which companies and how I can calculate the sentiment accordingly. Maybe a good labelled data-set or pre-trained architecture which might help me out?

P.S. Most important, is there any labelled dataset or any other pre-trained architecture which I can use to calculate the sentiments of financial news?

$\endgroup$
0
$\begingroup$

But how do I tell my model that it is actually a good news for Shell.

You would probably need a complex semantic analysis from which you could build a knowledge graph, from which you could extract the logical implications of one news for other entities not mentioned in the news.

P.S. Most important, is there any labelled dataset or any other pre-trained architecture which I can use to calculate the sentiments of financial news?

You could ask on https://opendata.stackexchange.com/. As far as I know, really good resources (data and algorithms) for financial applications are very expensive and usually kept secret by the biggest financial institutions.

| improve this answer | |
$\endgroup$
  • $\begingroup$ Thanks, makes sense! $\endgroup$ – Debadri Dutta Sep 25 '19 at 9:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.