Course Project for CMPE 239 - Allstate Purchase Prediction Challenge (Kaggle competition)

https://www.kaggle.com/c/allstate-purchase-prediction-challenge

https://github.com/farcryzry/spala/commit/cb126c86a4efe3c4ace6d824d889c3b915cf2563

https://github.com/farcryzry/spala/commit/2753caa7e2903135c6db6eb90e970217a3422069

Preprocess:

1. Filter out purchase records. (record_type = 1)
2. Remove columns: shopping_pt/record_type/day/time/state/location/c_previous/cost
3. Combine columns: age_oldest / age_youngest => age_average
4. Remove outlier: car_age
5. Handle missing value: Risk_factor (use 0) Duration_previous (use mean value)

Code: https://github.com/farcryzry/spala/tree/master/kmeans

Data: https://github.com/farcryzry/spala/tree/master/kmeans/data