Lists, lists and more lists…
Spam filter list, quality keyword list, stoplist list, top-domains urls list, news agencies websites list, university websites list, business websites lists and government organizations lists.
This gives us an initial score for the authority presenting the information.
If we can verify the source we are on the right track for building a fake news detector.
The spam filter also gives us clues on the quality of the source.
TESTING, TESTING and TESTING
Next step is running more tests to see if the concept works.
Then we need to find more lists and maybe other tools like an API that can give us clues im discovering fake news.
I don’t have all the answers but I am willing to code. It is a complicated problem and we may be limited on what can be done.
API may be a solution
I found two API that can make the project work.
URL Reputation API
With this URL Reputation API you can detect potentially phishing and malicious URLs. We deeply analyze the URL (including the URL content, URL pattern, domain name, HTTP headers, domain TLD, etc) It not free so I will abandonne the API for the moment.
I found another API that could help the project in a more complicated way.
Search API worldwide news https://newsapi.org/?ref=apilist.fun
We could cross reference news events with this API. We could us it to validate if the story is fake or is trending. But this could get complicated.
Memo Sim @ Fake news detector filters project
AFTER TESTING : THE LIST CONCEPT WORKS VERY WELL
The lists work very well together and the system is able to detect bad and good sites. I am very happy with this module. We are also able to get nice quality indicators and statistics for web page quality source evaluation.