Covid 19 Fact Checker

Given the current situation of the world there is an innumerable amount of fake news being publicized on various social media platforms. The intentions behind spreading these falsified news are various for eg:- political agenda, instilling fear in people’s mind, to grow ones business like selling of groceries or sanitizers, to gain money under the name of charity, etc.

• In order to eradicate this hoax we have come up with a system that can verify whether the news is fake or real.
• We intend to provide people sitting at home with real news, as news has become the only way to keep yourself updated.•Therefore in order to achieve our goal we have made use of 3 machine learning algorithms for classification which will be discussed further. 
• In addition to this we have generated a dataset comprising of news related to COVID-19. This data set, for time being, is generated by the team members of our group by referring to various sources like Times of India, Economic times, Facebook, Whatsapp, etc. This dataset set is generated in order to train our model.
• As our dataset is not clean, (which we have assumed will be the scenario when we are dealing with large data set with thousands of news) we have preprocessed it by using the following algorithms:-

1. Tf-idf Vectorization- term frequency–inverse document frequency, is a numerical statistic that is intended to reflect how important a word is to a document in a collection.
2. Count Vectorization-The CountVectorizer provides a simple way to both tokenize a collection of text documents and build a vocabulary of known words, but also to encode new documents using that vocabulary.

• Once we got the cleaned data set, we compared 3 classification algorithm :-

1. Multinomial Naïve Bayes
2. Decision tree Algorithm
3. Random forest

• We compared this on the basis of their accuracy and the result was that Multinomial Naïve Bayes gave the maximum accuracy among the three.
• Therefore we decided to classify the data using Multinomial NB, by using train test split we trained our model.Deployment :
• The platform used for deployment of our model was anvil where in we made an appealing UI where the user can check whether a news is true or fake.
• We have 3 approaches for verification of the news:-

1. Textbox:- News to be verified can be copied or typed manually.
2. Upload csv:- This version is included in our prototype where the user can upload a csv file containing news title and its description and the verification is made.
3. Upload image:- This is the extended version of text where our user can upload an image of a news, and with the help of OCR we will be able to verify. This version is our future goal where we are planning to make our model more flexible and advanced.

• In the time to come we are hoping to work with a larger data set to increase the accuracy of our model as more trained the dataset is more is the accuracy.