repeat-consumption

This is the code used for the paper Predicting Consumption Patterns with Repeated and Novel Events.

Data

The data used for this project can be found in the UCI Machine Learning repository here.

You can download the data and put in a ./data directory for the code to work as is.

If you want to change directory for where the data is loaded/saved you need to change util.paths file.

Reproducing Results

In order to reproduce the results from the paper you can run the python Notebook Experiments.

(The Data need to have been downloaded and be in the correct directory)

Using the Code

In order to use the Mixture Model described in the paper you can use the method train_mixture_model which is in models/mixture_model.py.

The method contains comments on how to use it.