Records Deleted, Target Missing: 1767

CART version 5.0.9.148

Records Read: 1930

Records deleted, target missing: 1767

Records Written in Learning sample: 163

Discrete N Levels

Variable in Model

------

HOLIDAY 2

Missing Value Prevalence

Learn

------

PRECIP 0.8344

STAB 0.0675

CURRENT MEMORY REQUIREMENTS

TOTAL: 18501. DATA: 1141. ANALYSIS: 18501.

AVAILABLE: 13500000. SURPLUS: 13481499.

The data are being read ...

163 Observations in the learning sample.

FILE: C:\SJV_CartAnalysis\cart_analysis\Bakersfield_holidays.txt

CART is running.

======

TREE SEQUENCE

======

Dependent variable: PM25

Terminal Cross-Validated Resubstitution Complexity Relative

Tree Nodes Relative Error Relative Error Parameter Complexity

------

1 30 0.552 +/- 0.074 0.177 0.000000 0.000

15* 11 0.433 +/- 0.058 0.279 1402.774048 0.011

16 10 0.448 +/- 0.058 0.294 1752.563110 0.014

17 9 0.448 +/- 0.053 0.311 2180.850098 0.018

18 8 0.485 +/- 0.055 0.331 2423.247559 0.020

19** 6 0.487 +/- 0.056 0.377 2777.187500 0.023

20 5 0.504 +/- 0.058 0.399 2797.292969 0.023

21 4 0.501 +/- 0.054 0.436 4453.742188 0.036

22 3 0.575 +/- 0.062 0.490 6671.226563 0.054

23 2 0.563 +/- 0.060 0.556 8070.929688 0.066

24 1 1.000 +/- 0.001 1.000 .543654E+05 0.444

Initial mean = 49.292

Initial variance = 751.477

======

NODE INFORMATION

======

*********************************************

* Node 1: STAB *

* N: 163 *

*********************************************

********************************* *********************************

* Node 2 * * Node 4 *

* N: 87 * * N: 76 *

********************************* *********************************

Node 1 was split on STAB

A case goes left if STAB <= 2.950

Improvement = 343.417999 Complexity Threshold = .543654E+05

Node Cases Wgt Count Mean StdDev

1 163 163.00 49.292 27.413

2 87 87.00 32.223 20.519

4 76 76.00 68.832 20.358

Surrogate Split Assoc Improvement

1 MINTEMP r 36.500 0.451 172.037

2 VIS r 3.500 0.268 198.118

3 RH r 55.850 0.183 4.969

4 PRECIP s 35.425 0.155 0.417

5 HOLIDAY s 0 0.127 63.607

Competitor Split Improvement

1 VIS 3.500 198.119

2 MINTEMP 36.500 172.037

3 HOLIDAY 0 63.607

4 RH 88.700 24.432

5 PRECIP 0.885 9.491

*********************************************

* Node 2: VIS *

* N: 87 *

*********************************************

********************************* ======

* Node 3 * = Terminal Node 3 =

* N: 32 * = N: 55 =

********************************* ======

Node 2 was split on VIS

A case goes left if VIS <= 2.500

Improvement = 49.514896 Complexity Threshold = 8070.929688

Node Cases Wgt Count Mean StdDev

2 87 87.00 32.223 20.519

3 32 32.00 44.850 25.783

-3 55 55.00 24.876 11.509

Surrogate Split Assoc Improvement

1 RH r 77.300 0.250 1.337

2 STAB s -10.160 0.063 2.402

3 MINTEMP s 30.500 0.031 4.789

4 HOLIDAY s 1 0.031 33.029

Competitor Split Improvement

1 STAB -1.985 34.352

2 HOLIDAY 0 33.029

3 MINTEMP 46.500 19.559

4 RH 63.650 5.579

5 PRECIP 0.885 3.773

*********************************************

* Node 3: RH *

* N: 32 *

*********************************************

======

= Terminal Node 1 = = Terminal Node 2 =

= N: 17 = = N: 15 =

======

Node 3 was split on RH

A case goes left if RH <= 79.450

Improvement = 27.323599 Complexity Threshold = 4453.742188

Node Cases Wgt Count Mean StdDev

3 32 32.00 44.850 25.783

-1 17 17.00 55.932 27.358

-2 15 15.00 32.291 16.522

Surrogate Split Assoc Improvement

1 VIS r 0.500 0.333 13.577

2 MINTEMP s 50.500 0.267 14.714

3 STAB r -3.155 0.267 23.004

4 PRECIP r 25.015 0.133 0.663

Competitor Split Improvement

1 STAB -4.430 25.655

2 HOLIDAY 0 22.959

3 MINTEMP 46.500 20.128

4 VIS 0.500 13.577

5 PRECIP 7.365 0.778

*********************************************

* Node 4: VIS *

* N: 76 *

*********************************************

********************************* ======

* Node 5 * = Terminal Node 6 =

* N: 46 * = N: 30 =

********************************* ======

Node 4 was split on VIS

A case goes left if VIS <= 2.500

Improvement = 40.927795 Complexity Threshold = 6671.226563

Node Cases Wgt Count Mean StdDev

4 76 76.00 68.832 20.358

5 46 46.00 76.398 19.382

-6 30 30.00 57.230 15.860

Surrogate Split Assoc Improvement

1 RH r 55.850 0.400 7.821

2 STAB r 8.965 0.233 11.662

3 MINTEMP s 40.500 0.100 13.084

Competitor Split Improvement

1 MINTEMP 46.500 19.087

2 STAB 8.760 13.969

3 RH 58.000 13.198

4 HOLIDAY 0 4.955

5 PRECIP 0.125 2.432

*********************************************

* Node 5: RH *

* N: 46 *

*********************************************

======

= Terminal Node 4 = = Terminal Node 5 =

= N: 37 = = N: 9 =

======

Node 5 was split on RH

A case goes left if RH <= 82.450

Improvement = 17.161308 Complexity Threshold = 2797.292969

Node Cases Wgt Count Mean StdDev

5 46 46.00 76.398 19.382

-4 37 37.00 80.244 17.203

-5 9 9.00 60.587 19.812

Surrogate Split Assoc Improvement

1 PRECIP r 0.125 0.111 2.432

Competitor Split Improvement

1 MINTEMP 47.000 6.870

2 STAB 12.430 4.952

3 VIS 0.500 3.110

4 PRECIP 0.125 2.432

5 HOLIDAY 0 0.020

======

TERMINAL NODE INFORMATION

======

Parent

Node Wgt Count Count Mean StdDev Complexity

------

1 17.00 17.000 55.932 27.358 4453.742

2 15.00 15.000 32.291 16.522 4453.742

3 55.00 55.000 24.876 11.509 8070.930

4 37.00 37.000 80.244 17.203 2797.293

5 9.00 9.000 60.587 19.812 2797.293

6 30.00 30.000 57.230 15.860 6671.227

======

VARIABLE IMPORTANCE

======

Relative Number Of

Importance Categories Penalty

------

STAB 100.000

VIS 79.409

MINTEMP 53.780

HOLIDAY 25.398 2

RH 15.404

PRECIP 0.923

N of the learning sample = 163

======

OPTION SETTINGS

======

Construction Rule Least Squares

Estimation Method 10-fold cross-validation

Tree Selection 1.000 se rule

Linear Combinations No

Initial value of the complexity parameter = 0.000

Minimum size below which node will not be split = 10

Node size above which sub-sampling will be used = 163

Maximum number of surrogates used for missing values = 5

Number of surrogate splits printed = 5

Number of competing splits printed = 5

Maximum number of trees printed in the tree sequence = 10

Max. number of cases allowed in the learning sample = 163

Maximum number of cases allowed in the test sample = 0

Max # of nonterminal nodes in the largest tree grown = 163

(Actual # of nonterminal nodes in largest tree grown = 34)

Max. no. of categorical splits including surrogates = 1000

Max. number of linear combination splits in a tree = 0

(Actual number cat. + linear combination splits = 9)

Maximum depth of largest tree grown = 16

(Actual depth of largest tree grown = 13)

Maximum size of memory available = 13500000

(Actual size of memory used in run = 63247)

======

CV-tree Competitor List

======

(Type, Predictor, Split if continuous, Improvement)

Top Split Competitors

Main 1 2 3 4

------

CV 1 | Numeric Numeric Numeric Categorical Numeric

| STAB MINTEMP VIS HOLIDAY RH

| 2.950 36.500 3.500 88.700

| 345.165 192.465 191.407 58.644 23.479

CV 2 | Numeric Numeric Numeric Categorical Numeric

| STAB VIS MINTEMP HOLIDAY RH

| 2.950 2.500 36.500 72.950

| 354.090 193.805 160.107 59.676 30.709

CV 3 | Numeric Numeric Numeric Categorical Numeric

| STAB VIS MINTEMP HOLIDAY RH

| 4.220 2.500 37.500 72.950

| 344.824 220.664 175.506 76.636 25.826

CV 4 | Numeric Numeric Numeric Categorical Numeric

| STAB VIS MINTEMP HOLIDAY RH

| 2.950 3.500 36.500 64.650

| 364.469 207.863 185.942 60.657 20.795

CV 5 | Numeric Numeric Numeric Categorical Numeric

| STAB VIS MINTEMP HOLIDAY RH

| 2.950 3.500 37.500 72.950

| 336.647 197.095 172.970 68.489 31.766

CV 6 | Numeric Numeric Numeric Categorical Numeric

| STAB MINTEMP VIS HOLIDAY RH

| 4.055 36.500 3.500 88.700

| 344.449 193.965 182.601 71.479 24.545

CV 7 | Numeric Numeric Numeric Categorical Numeric

| STAB VIS MINTEMP HOLIDAY RH

| 4.055 3.500 36.500 88.700

| 346.264 195.844 183.907 62.083 26.992

CV 8 | Numeric Numeric Numeric Categorical Numeric

| STAB VIS MINTEMP HOLIDAY RH

| 2.950 3.500 36.500 86.950

| 343.722 179.612 172.726 67.509 34.141

CV 9 | Numeric Numeric Numeric Categorical Numeric

| STAB VIS MINTEMP HOLIDAY RH

| 2.475 3.500 36.500 88.700

| 335.166 198.807 161.132 69.926 27.075

CV 10 | Numeric Numeric Numeric Categorical Numeric

| STAB VIS MINTEMP HOLIDAY RH

| 2.950 3.500 36.500 88.700

| 340.507 223.660 152.753 43.849 19.800

FINAL | Numeric Numeric Numeric Categorical Numeric

| STAB VIS MINTEMP HOLIDAY RH

| 2.950 3.500 36.500 88.700

| 343.418 198.119 172.037 63.607 24.432

Left Split Competitors

Main 1 2 3 4

------

CV 1 | Numeric Numeric Numeric Categorical Numeric

| VIS STAB MINTEMP HOLIDAY RH

| 2.500 -1.985 46.500 79.750

| 31.980 31.906 16.826 14.652 5.744

CV 2 | Numeric Categorical Numeric Numeric Numeric

| VIS HOLIDAY STAB MINTEMP RH

| 2.500 -1.985 46.500 63.650

| 43.511 37.016 33.614 18.665 10.013

CV 3 | Numeric Numeric Numeric Categorical Numeric

| STAB VIS MINTEMP HOLIDAY PRECIP

| -1.985 2.500 46.500 0.885

| 47.435 44.332 16.573 13.978 6.102

CV 4 | Numeric Numeric Categorical Numeric Numeric

| VIS STAB HOLIDAY MINTEMP PRECIP

| 2.500 -1.985 51.500 0.885

| 43.403 33.942 18.417 13.655 8.490

CV 5 | Numeric Numeric Categorical Numeric Numeric

| VIS STAB HOLIDAY MINTEMP RH

| 2.500 -2.130 46.500 79.450

| 49.439 41.214 35.724 28.270 8.117

CV 6 | Numeric Numeric Categorical Numeric Numeric

| VIS STAB HOLIDAY MINTEMP PRECIP

| 2.500 -1.930 46.500 0.885

| 45.715 42.129 32.838 27.408 7.682

CV 7 | Numeric Categorical Numeric Numeric Numeric

| VIS HOLIDAY STAB MINTEMP RH

| 2.500 -1.985 46.500 63.650

| 44.399 34.882 33.362 15.553 10.160

CV 8 | Numeric Categorical Numeric Numeric Numeric

| VIS HOLIDAY STAB MINTEMP RH

| 2.500 -1.985 46.500 79.450

| 50.364 41.901 33.837 22.595 8.122

CV 9 | Numeric Categorical Numeric Numeric Numeric

| VIS HOLIDAY STAB MINTEMP RH

| 2.500 -1.985 51.500 79.500

| 49.355 36.945 33.610 15.846 5.415

CV 10 | Numeric Numeric Categorical Numeric Numeric

| VIS STAB HOLIDAY MINTEMP RH

| 2.500 -3.985 46.500 63.350

| 65.276 33.094 30.825 20.928 4.497

FINAL | Numeric Numeric Categorical Numeric Numeric

| VIS STAB HOLIDAY MINTEMP RH

| 2.500 -1.985 46.500 63.650

| 49.515 34.352 33.029 19.559 5.579

Right Split Competitors

Main 1 2 3 4

------

CV 1 | Numeric Numeric Numeric Numeric Numeric

| VIS MINTEMP RH STAB PRECIP

| 2.500 36.500 58.000 8.495 1.265

| 61.439 33.093 15.603 14.921 12.788

CV 2 | Numeric Numeric Numeric Numeric Categorical

| VIS MINTEMP RH STAB HOLIDAY

| 2.500 46.500 58.000 8.965

| 44.851 23.031 15.643 15.170 4.994

CV 3 | Numeric Numeric Categorical Numeric Numeric

| VIS MINTEMP HOLIDAY RH STAB

| 3.500 47.000 43.450 11.030

| 66.064 16.147 12.589 11.489 10.675

CV 4 | Numeric Numeric Numeric Numeric Categorical

| VIS MINTEMP STAB RH HOLIDAY

| 3.500 36.500 8.760 58.000

| 41.667 18.765 16.036 13.826 6.147

CV 5 | Numeric Numeric Numeric Numeric Categorical

| VIS MINTEMP STAB RH HOLIDAY

| 3.500 36.500 4.055 43.450

| 41.035 18.211 11.919 9.263 7.229

CV 6 | Numeric Numeric Numeric Numeric Categorical

| VIS RH STAB MINTEMP HOLIDAY

| 3.500 44.050 12.530 47.500

| 40.299 13.939 11.653 10.506 3.156

CV 7 | Numeric Numeric Numeric Numeric Categorical

| VIS MINTEMP RH STAB HOLIDAY

| 3.500 45.500 52.750 12.530

| 34.411 19.094 15.570 12.178 3.640

CV 8 | Numeric Numeric Numeric Numeric Categorical

| VIS MINTEMP STAB RH HOLIDAY

| 3.500 36.500 8.760 58.000

| 35.751 21.406 13.809 13.448 3.700

CV 9 | Numeric Numeric Numeric Numeric Categorical

| VIS MINTEMP STAB RH HOLIDAY

| 2.500 46.500 12.530 58.000

| 34.890 20.504 15.459 11.360 9.223

CV 10 | Numeric Numeric Numeric Numeric Numeric

| VIS MINTEMP STAB RH PRECIP

| 2.500 46.500 8.760 58.000 0.125

| 45.005 20.550 17.767 17.165 2.697

FINAL | Numeric Numeric Numeric Numeric Categorical

| VIS MINTEMP STAB RH HOLIDAY

| 2.500 46.500 8.760 58.000

| 40.928 19.087 13.969 13.198 4.955

======

Gains for PM25

======

Mean P of Cum P of Cum

Target Target P of N Pop Cum P Lift Lift

Node in Bin in Bin Target in Bin in Bin of Pop Ratio Index

------

3 81.7233353 0.0305 0.0305 3.00 0.0184 0.0184 1.66 1.66

7 80.2440491 0.3695 0.4000 37.00 0.2270 0.2454 1.63 1.63

9 74.6672668 0.1022 0.5023 11.00 0.0675 0.3129 1.61 1.51

2 67.1500015 0.0418 0.5441 5.00 0.0307 0.3436 1.58 1.36

8 60.5866699 0.0679 0.6119 9.00 0.0552 0.3988 1.53 1.23

10 50.7350006 0.0631 0.6751 10.00 0.0613 0.4601 1.47 1.03

11 43.1355515 0.0483 0.7234 9.00 0.0552 0.5153 1.40 0.88

1 41.1022224 0.0460 0.7694 9.00 0.0552 0.5706 1.35 0.83

4 32.2906609 0.0603 0.8297 15.00 0.0920 0.6626 1.25 0.66

6 31.5343437 0.0903 0.9200 23.00 0.1411 0.8037 1.14 0.64

5 20.0903111 0.0800 1.0000000000 32.00 0.1963 1.0000 1.00 0.41

------

163.00

C:\DOCUME~1\Rhackney\LOCALS~1\Temp\s34865: 20.9 kb

Grove file created containing 1 Tree.

C:\SJV_CartAnalysis\cart_analysis\Bakersfield_holidays.txt: 2135 records (estimated).