Group 28 - MyGenre: A text-based content recommendation application

Group Members

Anna Marion, Dan Rieman, Meghana Srivatsa

Abstract

MyGenre is a text-based application that generates recommendations based upon user input. It leverages content from two domains: social media and movies. The application structure calculates linguistic similarities among texts. It then compares word frequencies to classify individual titles or users as a specific genre. Users can enter text and receive lexically similar movies or Twitter accounts based upon cosine similarities of vector representations. Users can also obtain wordclouds as a visual measure of word frequency.

Pictures