Open the attached CSV file containing the share prices for the stock from Tesla Inc. (TSLA). The file contains seven (7) columns:
- Opening Price for the Day
- Highest Price for the Day
- Lowest Price for the Day
- Closing Price for the Day
- Adjusted Closing Price
- Trading Volume
This data has been gathered over a 7 year period between 2010 and 2017.
Your assignment is to forecast the values for the week of the last date of the stock listing. Given the last share price listing is for March 7, 2017, which was a Tuesday, use the given data to predict the values for:
a) Opening Price, b) Closing Price, c) Highest Price, d) Lowest Price and, e) Trading Volume
for that entire week beginning Monday, Mar 6, 2017, and ending Friday, March 10, 2017
STEP 1. Start with using the simple average window method to forecast the prices for the previous week, i.e, the week starting Monday, Feb 27, 2017, and ending Friday, March 3, 2017. For the width of the window, use values of k = 2, 3, and 4.
STEP 2. Use a weighted average window method with at least 2 different weight combinations to forecast the prices for the week, starting Monday, February 27, 2017, and ending Friday, March 3, 2017. Use a window width of k = 3. Remember, the sum of weights should equal to 1.0
STEP 3: Compare the performance of your average windows methods in STEP 1 and STEP 2 using the three different error methods discussed in class: Mean Absolute Deviation (MAD), Mean Squared Error (MSE), and Mean Average Percentage Error (MAPE). Note your comparison must include at least five different columns: three for the simple moving average window and at least two for the weighted average window. On the basis of your comparison, which method has the best performance?
STEP 4: Repeat Steps 1~3 using the exponential smoothing method. Choose at least 3 values for the smoothing parameter Alpha. Does the performance of your forecasting method improve when using the exponential smoothing method?