The problem NewsFresh solves

Fake News has been a modern-day challenge for many people on Social Media and there have been very fewer efforts on tackling it on Social Media Platforms where people are more subservient (not ready to accept other unquestionably) towards Fake News. This has led to many hate campaigns now being organized in the name of “facts” which usually sprout out from dubious news sources. To battle this, multiple kinds of research have been carried out in past which can help detect fake news using statistical techniques, which today has assumed the form of Machine Learning and Artificial Intelligence.

However, when we further went down the research, we found that Machine Learning and Artificial Intelligence cannot be a one-stop solution to this ever-increasing problem. This has led us to adopt a Cognitive approach towards the problem which not only relies on statistical techniques but also logical reasoning by crowdsourcing the Machine Learning Model to help gather data from multiple users on a Web/Android/Chromium Platform which will allow us to perfect our model.

Challenges we ran into

We ran into a serious problem when we started to scrap News Articles from all over the Web. The primary shortcoming of our approach was that even unrelated news articles were scrapped which was seriously affecting our reputabilty algorithm. To tackle this, we implemented a Cosine Similarity Algorithm which can check out related news articles. We also ran into UI/UX Bugs which were eventually fixed.