In: Statistics and Probability
ASAP!!!!!!!!!
The monthly sales for Yazici Batteries, Inc., were as follows: Month Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec Sales 21 21 15 15 11 18 16 18 19 20 23 22 This exercise contains only parts b and c.
b) The forecast for the next month (Jan) using the naive method = 22 sales (round your response to a whole number). The forecast for the next period (Jan) using a 3-month moving average approach sales (round your response to two decimal places).
The forecast for the next period (Jan) using a 6-month weighted average with weights of 0.10, 0.10, 0.10, 0.20, 0.20, and 0.30, where the heaviest weights are applied to the most recent month = sales (round your response to one decimal place).
Using exponential smoothing with alpha = 0.30 and a September forecast of 21.00, the forecast for the next period (Jan) = nothing sales (round your response to two decimal places).
Using a method of trend projection, the forecast for the next month (Jan) = nothing sales (round your response to two decimal places).
c) The method that can be used for making a forecast for the month of March is ▼ a 3-month moving average a 6-month weighted moving average exponential smoothing the naive method a trend projection .
b) The forecast for the next month (Jan) using the naive method = 22 sales (round your response to a whole number).
In the naive method, we simply believe that the current forecast will be the most recent trend.
So for Feb forecast it will Jan's actual data.
Naïve method | ||
Month | Actual | Forecast |
Jan | 21 | |
Feb | 21 | 21 |
Mar | 15 | 21 |
Apr | 15 | 15 |
May | 11 | 15 |
Jun | 18 | 11 |
Jul | 16 | 18 |
Aug | 18 | 16 |
Sep | 19 | 18 |
Oct | 20 | 19 |
Nov | 23 | 20 |
Dec | 22 | 23 |
Jan (+1) | 22 |
The forecast for the next period (Jan) using a 3-month moving average approach sales (round your response to two decimal places).
For this we first take the sum of most recent 3 period. Then take their average. This is the forecast for the 4th period
Eg: For May = (Feb + Mar + Apr) / 3
if it was 4 period then we would take totals of 4 periods
3-period average | |||
Month | Actual | 3 period sum | Average |
Jan | 21 | ||
Feb | 21 | ||
Mar | 15 | ||
Apr | 15 | 57 | 19 |
May | 11 | 51 | 17 |
Jun | 18 | 41 | 13.667 |
Jul | 16 | 44 | 14.667 |
Aug | 18 | 45 | 15 |
Sep | 19 | 52 | 17.333 |
Oct | 20 | 53 | 17.667 |
Nov | 23 | 57 | 19 |
Dec | 22 | 62 | 20.667 |
Jan (+1) | 65 | 21.667 |
The forecast for the next period (Jan) using a 6-month weighted average with weights of 0.10, 0.10, 0.10, 0.20, 0.20, and 0.30, where the heaviest weights are applied to the most recent month = sales (round your response to one decimal place).
For weighted average we assign different weights to different periods
This is 6 period, so the forecast will alwys be for next 7th period.
Heavier to recent weight means if for July, most recent will be June then gradually going upto Jan
July = 0.1 *(Jan + Feb + Mar) + 0.2 (Apr + May) + 0.3 * June
6-period weighted average | ||
Month | Actual | Average |
Jan | 21 | |
Feb | 21 | |
Mar | 15 | |
Apr | 15 | |
May | 11 | |
Jun | 18 | |
Jul | 16 | 16.3 |
Aug | 18 | 15.7 |
Sep | 19 | 16.3 |
Oct | 20 | 16.9 |
Nov | 23 | 17.9 |
Dec | 22 | 19.9 |
Jan (+1) | 20.5 |
Using exponential smoothing with alpha = 0.30 and a September forecast of 21.00, the forecast for the next period (Jan) = nothing sales (round your response to two decimal places).
Here the exponential smoothing gives two weights to previous actual and forecast values. Heavier is to the forecast
= 0.3
Ft = At-1 * + (1 - ) Ft-1
Eg; Here we begin with Sep since its forecast is given
October = 19 * 0.3 + 21 * 0.7
Exponenetial | ||
Month | Actual | Forecast |
Jan | 21 | |
Feb | 21 | |
Mar | 15 | |
Apr | 15 | |
May | 11 | |
Jun | 18 | |
Jul | 16 | |
Aug | 18 | |
Sep | 19 | 21 |
Oct | 20 | 20.4 |
Nov | 23 | 20.28 |
Dec | 22 | 21.096 |
Jan (+1) | 21.37 |
Using a method of trend projection, the forecast for the next month (Jan) = nothing sales (round your response to two decimal places).
We can use the regression method for this. Where the 'x' is the year. And y is the sales. The 'x' can take values from 1 - 12 representing the number of month. We find a regression equation to predict the future values.
Month (x) | Sales (y) | x^2 | y^2 | xy |
1 | 21 | 441 | 1 | 21 |
2 | 21 | 441 | 4 | 42 |
3 | 15 | 225 | 9 | 45 |
4 | 15 | 225 | 16 | 60 |
5 | 11 | 121 | 25 | 55 |
6 | 18 | 324 | 36 | 108 |
7 | 16 | 256 | 49 | 112 |
8 | 18 | 324 | 64 | 144 |
9 | 19 | 361 | 81 | 171 |
10 | 20 | 400 | 100 | 200 |
11 | 23 | 529 | 121 | 253 |
12 | 22 | 484 | 144 | 264 |
Total | 219 | 4131 | 650 | 1475 |
Mean | 18.25 | 344.25 |
Regression eq of Y on X
slope =
= 0.360
intercept
=15.909
The regression equation is
= 15.909 + 0.36x
We sub Jan = 13th month so x = 13
= 15.909 + 0.36 * 13
= 20.591
Linear regression helps to predict equation of line where when the x,y points are plotted they seem to form a linear line.
c) The method that can be used for making a forecast for
the month of March is ▼ a 3-month moving average a 6-month weighted
moving average exponential smoothing the naive method a trend
projection
Method | Value of march |
Naïve | 21 |
moving avg | - |
Weighted | - |
Exponential | - |
linear | 16.98951 |
for linear trend we sub x = 3 in the regression equation.
So the methods can be used to foreast MArch are Naive and Trend projection.