In: Statistics and Probability
The number of internal disk drives (in millions) made at a plant in Taiwan during the past 5 years follows: Year 1 2 3 4 5 Disk Drives 138 160 196 200 210 a) Using simple linear regression LOADING... , the forecast for the number of disk drives to be made next year = nothing disk drives (round your response to one decimal place). b) The mean squared error (MSE) when using simple linear regression = nothing drives squared (round your response to one decimal place). c) The mean absolute percentage error (MAPE) when using simple linear regression = nothing % (round your response to one decimal place).
| Year | Disk Drives | 
| 1 | 138 | 
| 2 | 160 | 
| 3 | 196 | 
| 4 | 200 | 
| 5 | 210 | 
a.) Linear regression is of the form
y = a + b*x
First finding mean, standard devaition for both x and y we get
mx = 15/5 = 3
my = 904/5 = 180.8
sx = 1.58
sy = 30.482
Correlation between x and y is given as,
r = 0.954
So, b = r * (sy/sx)
b = 0.954 *(30.482/1.58)
b = 18.4
a = my - b*mx
a = 180.8 - 18.4 * 3
a = 125.58 = 125.6
So, equation becomes,
y(disk drive) = 125.6+ 18.4 * x(year)
b.)
Mean Squared Error = Sum of squared error / n
First predtcing values of y based on above equation,
| Year | Disk Drives | Predicted | (Error)^2 | 
| 1 | 138 | 144 | 36 | 
| 2 | 160 | 162.4 | 5.76 | 
| 3 | 196 | 180.8 | 231.04 | 
| 4 | 200 | 199.2 | 0.64 | 
| 5 | 210 | 217.6 | 57.76 | 
Sum of squared error = 331.2
So, Mean Squared error = 331.2 / 5 = 66.24
c.)

If we multiply above by 100, we get percentage.
| Year | Disk Drives | Predicted | |Actual - Predicted| | |Actual - Predicted| / Actual | 
| 1 | 138 | 144 | 6 | 0.0435 | 
| 2 | 160 | 162.4 | 2.4 | 0.0150 | 
| 3 | 196 | 180.8 | 15.2 | 0.0776 | 
| 4 | 200 | 199.2 | 0.8 | 0.0040 | 
| 5 | 210 | 217.6 | 7.6 | 0.0362 | 
Sum of Last column = 0.1762
So, MAPE = (0.1762/5) * 100
MAPE = 3.524