Home Automation Analytics

Data for recommender system: Jester5k dataset from CRAN's recommenderlab. Data for predictive maintenance system: Created.

Purpose for project:

Representatives from IBM Japan hired the Alacer Group, LLC. for home automation predictive analytics in a marketing context. The Statistician role involved creating a predictive maintenance and recommender system demo for home automation machines. I developed test data for the predictive maintenance statistics models and a data mining recommender system as a 1:1 marketing application for suggesting warranty parts to end-users. Data for machine breakdowns come in next year, so not results at this time.

Algorithms:

Recommender system used User-based and Item-based Collaborative Filtering.

Predictive maintenance system used Logistic regression classifier.