The code includes the starter code provided at:

http://www.kaggle.com/c/belkin-energy-disaggregation-competition/download/LoaderScripts.zip

This was one of my first data mining projects. The main challenges faced were the size and structure of data. To overcome these problems, the data was downsampled and redundant data was removed. Also, all operations had to be made vectorized.

The KNN algorithm was used to train and predict the disaggregated consumption of households.

The dataset for the project can be found on the kaggle website mentioned in the description. LoadData.m should be used to run the project.