by @jaimemorillo

Automatic Prediction of Affinity to a Movie from the Synopsis

In this repository I will place my TFG (Thesis) Software application, which helps people to find interesting movies in a different way to traditional recommendation systems.

Contact

• E-mail: jamorillo.leal@gmail.com

• Github: Jaimemorillo

About the project

Topic

Web application that shows you movies and their scoring based on your personal taste and an historic of previous movies seen.

How is it done

Solving an NLP Binary Classification problem training a neural network using the synopsis of the movies (in Spanish).

Techniques

The following nlp techniques have been applied:

• Normalization and erasing of stopwords.

• Stemming.

• Tokenization.

• Word2Vec Emmbedings.

• CNN and RNN.

How to use

• To run the app:

  python app/app.py

• Go to http://127.0.0.1:5000/

What can you do?

• Movie labeling.

• Read the synopsis.

• See the posters.

• See random movies generated.

• See the actual Billboard.

• Automatic model training and prediction (real time).

Demo