Objective

Predict the weekly sales for the subsequent two months for every combination of Store, Dept, and Date in test.csv using the historical data from train.csv.

• train.csv contains 5 columns ("Store", "Dept", "Date", "Weekly_Sales", "IsHoliday") and ranges
  • from 2010-02 (February 2010) to 2011-02 (February 2011) in fold_1,
  • from 2010-02 to 2011-04 in fold_2,
  • from 2010-02 to 2011-06 in fold_3,
  • ......
  • from 2010-02 to 2012-08 in fold_10.
• test.csv contains 4 columns ("Store", "Dept", "Date", "IsHoliday") and ranges
  • from 2011-03 to 2011-04 in fold_1,
  • from 2011-05 to 2011-06 in fold_2,
  • from 2011-07 to 2011-08 in fold_3,
  • ......
  • from 2012-09 to 2012-10 in fold_10.
• test_with_label.csv is formatted similarly to train.csv and ranges from 2011-03 to 2012-10.

Here is data set : https://liangfgithub.github.io/Data/proj2.zip

Evaluation This competition is evaluated on the weighted mean absolute error (WMAE): WMAE = M= wilyi - yil E wi 1=1 where . n is the number of rows . \\\\(\\hat(y)_i \\\\ is the predicted sales . \\\\(y_i \\\\) is the actual sales . \\\\(w_i \\\\) are weights. w = 5 if the week is a holiday week, 1 otherwise