Inspiration
Given the current virus crisis and a lack of epistemic access to frontline research, millions of people around the world currently face a crisis in terms of determining the fake news from the real. With cases soaring all around the world, its crucial at this juncture to enable quick access to people who want to determine the efficacy of their news sources.
What it does
We have used the fake news dataset from kaggle and have trained a model structured as a Decision Tree Classifier. We have made a webpage where users can enter in the sentences they want to check and obtain instantaneous results.
How we built it
We used google colab to build and train the ML model. The frontend is primarily built on html5, css, js and jQuery libraries. The help of some web flow scripts are taken to ensure a better looking frontend. We have chose flask as the backend framework. It renders the html page and the news to be verified it POST to the backed where it's processed using the ML algorithm. Depending on the result the output is returned.
Challenges we ran into
Integrating the front end with the backend especially using heroku to install all dependencies caused us unexpected delays in deploying the entirety of the web app.
Accomplishments that we're proud of
We got extremely high accuracy percentages out of our model and got a smooth flow UI established on our site.
What we learned
How to integrate colab notebooks into the backend of a site and building a smooth flowing UI for the front end.
Log in or sign up for Devpost to join the conversation.