Inspiration
Social Media Approval machine learning challenge at TAMU Datathon
What it does
Predicts whether or not a social media post is approved based on its caption
How we built it
Using google collab, we used tensorflow and panda to create a sequential model that pulls the given data, tokenizes each caption, and predicts a pattern of approval based on those captions.
Challenges we ran into
The model often overfits itself to the training data
Accomplishments that we're proud of
Learning how to build a machine learning model from the ground up with little to no experience in machine learning
What we learned
We learned computer vision, natural language processing, and webscrapping from workshops, as well as what processes are used to build and run a sequential model.
What's next for Social Media Approval
Next is to incorporate other variables into the model in order to consider other points besides caption data in social media posts without lowering the validation accuracy of the model as well as create a way to prevent overfitting of the training data.
Log in or sign up for Devpost to join the conversation.