Net Trained on AWEX 2007 Jul-Sep PCS modifiers
Selling centre Fremantle
Sale week W10
Style Good
Diameter 19.5
Yield 50.0
Vegetable Matter 1.0
Staple Length 85
Staple Strength 25
Hauteur 57
Clean price (cents per kg) #NAME?
SL SS VMB
50 15 0.5
55 20 1.0
60 25 1.5
65 30 2.0
70 35 3.0
75 40 4.0
80 45 5.0
85 50
90 55
95 60
100 65
105 70
110
85 25 1.0
micron Yield team2H Centre Week No saleoutcome Style MID
17.0 40 35 Fremantle W01 passed in Best 10
17.5 45 40 Melbourne W05 Sold Good 20
18.0 50 45 Sydney W06 Traded Average 30
18.5 55 50 W07 40
19.0 60 55 W08 50
19.5 65 60 W09 60
20.0 70 65 W10 70
20.5 75 70 W11 80
21.0 75 W12 90
21.5 100
22.0
22.5
23.0
23.5
24.0
F PI 4
M S 5
S TR 6
19.5 50 57 F W10 S 5 #NAME?
Tag SELLINGCENTRE SaleWeekNum SALEOUTCOME SL SS VMB style micron yield cprice team2H
predict F W10 S 85 25 1 5 19.5 50 #NAME? 57
The training and testing data
has been stripped out
of this workbook
to stream-line computation
and to reduce file size.
NeuralTools (Report: Neural Net Training and Auto-Prediction)
Created for: Kimbal Curtis
Date: Monday, 15 October 2007
Summary
Net Information
Name Net Trained on AWEX 2007 Jul-Sep PCS calculator
Configuration GRNN Numeric Predictor
Location FMS Postsale data_Jul-Sep_2007.v1b.calculator.xls
Independent Category Variables 4 (SELLINGCENTRE, SaleWeekNum, SALEOUTCOME, style)
Independent Numeric Variables 6 (SL, SS, VMB, micron, yield, team2H)
Dependent Variable Numeric Var. (cprice)
Training
Number of Cases 6908
Training Time (h:min:sec) 0:35:09
Number of Trials 74
Reason Stopped Auto-Stopped
% Bad Predictions (10% Tolerance) 2.0122%
Root Mean Square Error 28.43
Mean Absolute Error 19.38
Std. Deviation of Abs. Error 20.80
Prediction
Number of Cases 1
Live Prediction Enabled YES
Data Set
Name AWEX 2007 Jul-Sep PCS calculator
Number of Rows 6911
Manual Case Tags YES
Variable Impact Analysis
Not Completed Analysis Stopped by User
Histogram of Residuals (Training)
3000
2500
2000
Frequency
1500
1000
500
0
-50.00
0.00
50.00
-250.00
-200.00
-150.00
-100.00
100.00
150.00
200.00
Predicted vs. Actual (Training)
2500
2000
1500
Predicted
1000
500
0
0 500 1000 1500 2000 2500
Actual
Residual vs. Actual (Training)
200
150
100
50
Residual
0
0 500 1000 1500 2000 2500
-50
-100
-150
-200
-250
Actual
Residual vs. Predicted (Training)
200
150
100
50
Residual
0
0 500 1000 1500 2000 2500
-50
-100
-150
-200
-250
Predicted