KentonParton/game_revenue

Using machine learning and feature data of past video games, the model can predict the potentail revenue generated by a particular game.

game_revenue

Requirements

tensorflow==1.1.0

keras==2.1.5

pandas==0.22.0

Files

scale.py

Scales data from .csv file before training.

train_model.py

Trains model using scaled data and creates Logs which can be used on Tensorboard and trained_model.h5.

predict.py

Uses trained_model.h5 to predict potential revenue a game will earn based on proposed_new_product.csv

To view logs in Tensorboard, tensorboard --logdir=game_revenue/logs