Problem
Novine knows that over the last 10 years average computer prices have fallen. She believes that average computer prices can be predicted by time, computing speed (number of instructions processed per second), memory (RAM), storage (hard drive space), and processor chip price.
1. Using Excel, develop a full linear regression model to predicted the average computer price using the variables above. Based on the model, rank the variables (using p-values) based on degrees of significance in the model.
2. Calculate the MAD for your model.
3. Observe R-squared, Significant-F, and t statistics and interpret them.
4. What is the best two variable regression model to predict average computer price? Is there a problem with using linear regression on this data?
|
Avg computer price
|
Year
|
Bits
|
RAM
|
Storage
|
Chip Price
|
|
$ 3,000.00
|
1990
|
2
|
4
|
40
|
870
|
|
$ 2,850.00
|
1991
|
4
|
8
|
80
|
800
|
|
$ 2,750.00
|
1992
|
4
|
8
|
80
|
790
|
|
$ 2,250.00
|
1993
|
8
|
16
|
100
|
700
|
|
$ 2,100.00
|
1994
|
8
|
24
|
200
|
500
|
|
$ 1,900.00
|
1995
|
16
|
32
|
500
|
450
|
|
$ 1,850.00
|
1996
|
16
|
64
|
500
|
400
|
|
$ 1,700.00
|
1997
|
32
|
64
|
1000
|
375
|
|
$ 1,500.00
|
1998
|
64
|
128
|
2000
|
350
|
|
$ 1,100.00
|
1999
|
64
|
128
|
4000
|
300
|