Problem 1: Forecasting (20 points) Consider the weekly sales dataset included. This file contains the following columns:

Date Weekly Sales Holiday: this column is 1 if the date involves a holiday, and 0 otherwise

Temperature

Fuel Price

CPI: consumer price index; provides a measure for inflation

Unemployment: percentage of unemployment Predictive Analysis Considering the information described above, your goal is to forecast weekly sales. To do so, use the following methods:

1. Nave Forecast

2. Moving Average, with N=3

3. Weighted Average, with weights !"# = 0.5, !"$ = 0.3, !"% = 0.2.

4. Exponential Smoothing, with smoothing constant 0.5

5. Linear regression (considering all predictor variables) For each method, do the following:

Compute your forecast Plot the actual data and your forecast Compute the MAE, MSE and MAPE. Based on your forecasts and the measures of goodness of fit, which forecast method is better?

Date	Weekly_Sales	Holiday	Temperature	Fuel_Price	CPI	Unemployment
2/5/2010	24924.5	0	42.31	2.572	211.0963582	8.106
2/12/2010	46039.49	1	38.51	2.548	211.2421698	8.106
2/19/2010	41595.55	0	39.93	2.514	211.2891429	8.106
2/26/2010	19403.54	0	46.63	2.561	211.3196429	8.106
3/5/2010	21827.9	0	46.5	2.625	211.3501429	8.106
3/12/2010	21043.39	0	57.79	2.667	211.3806429	8.106
3/19/2010	22136.64	0	54.58	2.72	211.215635	8.106
3/26/2010	26229.21	0	51.45	2.732	211.0180424	8.106
4/2/2010	57258.43	0	62.27	2.719	210.8204499	7.808