UnFake-IT

 

 Do you know that Virat Kohli has retired from International Cricket? Or that Britain is planning to re-enter the Elections?

If you are tired of these mind-numbing news that you know are fake but have to believe and bear it because it is everywhere and you don't have the time and energy to verify it!

So, then we, the, TEAM AMALGAM present you with the perfect tool that provides a perfect inferring that classifies news by colour grading the news post and giving the viewer how credible the report is.

UNFAKE-IT 

BRAIN-STORMING

As we were provided with the  given example given with the associated statement that how the day after demonitization was announced,there was a viral picture that reported that the new Rs.2000 note has a nano GPS.

So we pondered the root cause of all these viral fake news is ,that anybody and everybody could anything at anytime ,there is little or no viable sources to pace to get our facts right on social media platforms whether it be facebook,twitter ,quora & reddit etc.

So either we can only let reputable,genuine and trustworthy news broadcaster to publish there post or we can grade the normal posts published by the general public.

 

Clearly the second option seems more viable as we cant deny anyone there fundamental freedom to right of expression.So to grade the post we devised a basic algorithm of mapping  and matching the general post with the known trust worthy ones.

 

So we have developed and designed a social media platform plugin or web browser extension to deal with this problem statement.

ALGORITHM FORULATED

We devised an unique algo for classifying and grading our posts by making it compulsory that everytime whenever a user uploads any posts ,he/she has to attach a min of suggestive tags and a max of 15.

The tags would be automatically generated according to the content ,context and the present location of the user to give him relevant and coordinated tags to the user to th user in drop down list consisting of n no of options in the form of checkboxes.

So when the user hits the post button using ML techniques our tags will help us in associating a main category to the post associated,that main category would have many subcategories to form a perfect query for a given post.

This will all be analysed in real time and tree structure of tags -category-subcategory will provide  unique status to our post acc to which we can associate a genuine new source(government website-major tabloids for verification and inferencing purposes).

Each post will have a set of sensitive(high priority) and non sensitive(low priority) that will b automatically identified by our algo itself.The same set of keywords are matched and mapped to source website's article to calculate its relevance score.

Acc to its numeric score each post would be colour graded above the content mentioned in it. Where green being the highest in genuineness ,yellow being a bt ambiguous ,orange being difficult to make out and red being almost fake.

Ex. consider the 2000  rupees nano gps chip fake news.It would have tag like demonitization.Acc. to this we would assign it main category of finance and sub category of money and its genuine source would be RBI website .

So when post is published it would be color graded and ready to be infered to viweing by public which can either support or oppose grading assigned to the given post if we get  a requisite number of oppose votes than we would downgrade its current color grading by one step.This would be follow for each iteration. This would help in those cases where the source could  be some information to cover up its own fallacies.

Whenever there is any random post which has no relevance to current scenario of events we would assign it a black label.

                      Surveys Analysis