Initial Experiments

Intro ] [ shorttest1 ] [ simpletest-no-missing ] [ simpletest1 ] [ test1 ] [  ]

Up: Report ]

Naive Bayes (simple)

Class >50K: P(C) = 0.22709163

Attribute age
Mean: 42.92920354	Standard Deviation: 9.86037961

Attribute workclass
Private		Self-emp-not-inc	Self-emp-inc	Federal-gov
0.58119658	0.08547009		0.08547009	0.07692308
Local-gov	State-gov		Without-pay	Never-worked	
0.06837607	0.08547009		0.00854701	0.00854701


Attribute fnlwgt
Mean: 188417.6460177	Standard Deviation: 97710.42446423

Attribute education
Bachelors	Some-college	11th		HS-grad		Prof-school
0.24031008	0.17054264	0.00775194	0.17829457	0.04651163
Masters		1st-4th		10th		Doctorate	5th-6th
0.04651163	0.0620155 	0.00775194	0.02325581	0.01550388
Assoc-acdm	Assoc-voc	9th		7th-8th		12th
0.11627907	0.00775194	0.00775194	0.05426357	0.00775194
Preschool	
0.00775194


Attribute education-num
Mean: 11.66371681	Standard Deviation: 2.40390538

Attribute marital-status
Married-civ-spouse	Divorced		Never-married
0.825			0.075			0.06666667
Separated		Widowed			Married-spouse-absent
0.00833333		0.00833333		0.00833333
Married-AF-spouse	
0.00833333	


Attribute occupation
Tech-support		Craft-repair		Other-service	Sales
0.04878049		0.11382114		0.04065041	0.08943089
Exec-managerial		Prof-specialty		Handlers-cleaners	
0.26829268		0.21138211		0.00813008
Machine-op-inspct	Adm-clerical		Farming-fishing	
0.04878049		0.07317073		0.02439024
Transport-moving	Priv-house-serv		Protective-serv
0.01626016		0.00813008		0.04065041
Armed-Forces	
0.00813008	

Attribute relationship
Wife		Own-child	Husband		Not-in-family	Other-relative
0.13445378	0.00840336	0.69747899	0.11764706	0.01680672
Unmarried
0.02521008	

Attribute race
White		Asian-Pac-Islander	Amer-Indian-Eskimo	Other
0.83050847	0.07627119		0.00847458		0.00847458
Black	
0.07627119

Attribute sex
Female		Male	
0.20869565	0.79130435	

Attribute capital-gain
Mean: 1827.6460177	Standard Deviation: 4200.9996534

Attribute capital-loss
Mean: 202.79646018	Standard Deviation: 594.83464328

Attribute hours-per-week
Mean: 44.38053097	Standard Deviation: 10.95471081

Attribute native-country
United-States	Cambodia	England		Puerto-Rico	Canada
0.66666667	0.01333333	0.00666667	0.00666667	0.01333333
Greece		South		China		Cuba		Iran
0.00666667	0.00666667	0.01333333	0.00666667	0.00666667
Honduras	Philippines	Italy		Poland		Jamaica
0.01333333	0.00666667	0.00666667	0.02      	0.01333333
Vietnam		Mexico		Portugal	Ireland		France
0.00666667	0.00666667	0.00666667	0.00666667	0.00666667
Dominican-Republic		Laos		Ecuador		Taiwan
0.01333333			0.00666667	0.00666667	0.00666667
Haiti		Columbia	Hungary		Guatemala	Nicaragua
0.00666667	0.00666667	0.00666667	0.01333333	0.00666667
Scotland	Thailand	Yugoslavia	El-Salvador	Trinadad&Tobago
0.00666667	0.00666667	0.00666667	0.00666667	0.00666667
Germany		Outlying-US(Guam-USVI-etc)	India		Japan
0.01333333	0.00666667			0.00666667	0.00666667
Peru		Hong		Holand-Netherlands	
0.00666667	0.00666667	0.00666667


Class <=50K: P(C) = 0.77290837

Attribute age
Mean: 36.54005168	Standard Deviation: 14.0142534

Attribute workclass
Private		Self-emp-not-inc	Self-emp-inc	Federal-gov
0.76358696	0.07336957		0.01630435	0.02717391
Local-gov	State-gov		Without-pay	Never-worked	
0.07608696	0.03804348		0.00271739	0.00271739	

Attribute fnlwgt
Mean: 197706.94573643	Standard Deviation: 119881.03688158

Attribute education
Bachelors	Some-college	11th		HS-grad		Prof-school
0.12406948	0.23573201	0.05955335	0.33746898	0.00248139
Assoc-acdm	Assoc-voc	9th		7th-8th		12th
0.05210918	0.03473945	0.0248139 	0.01736973	0.00496278
Masters		1st-4th		10th		Doctorate	5th-6th
0.04218362	0.01488834	0.0248139 	0.00248139	0.01736973
Preschool	
0.00496278	

Attribute education-num
Mean: 9.6124031 	Standard Deviation: 2.47152722

Attribute marital-status
Married-civ-spouse	Divorced		Never-married	Separated
0.34010152		0.16497462		0.40862944	0.02791878
Widowed			Married-spouse-absent	Married-AF-spouse	
0.02791878		0.02538071		0.00507614	

Attribute occupation
Tech-support		Craft-repair	Other-service
0.04278075		0.14438503	0.14438503	
Sales			Exec-managerial	Prof-specialty	Handlers-cleaners
0.13368984		0.07219251	0.09625668	0.05882353
Machine-op-inspct	Adm-clerical	Farming-fishing	Transport-moving
0.06417112		0.11229947	0.02941176	0.07219251
Priv-house-serv		Protective-serv	Armed-Forces	
0.0026738	 	0.02139037	0.00534759	

Attribute relationship
Wife		Own-child	Husband		Not-in-family	Other-relative
0.05343511	0.18320611	0.28244275	0.33587786	0.03816794
Unmarried	
0.10687023	

Attribute race
White		Asian-Pac-Islander	Amer-Indian-Eskimo	Other
0.81632653	0.0255102	 	0.0127551	 	0.0127551
Black	
0.13265306	

Attribute sex
Female		Male	
0.37017995	0.62982005	

Attribute capital-gain
Mean: 195.86563307	Standard Deviation: 1824.06242721

Attribute capital-loss
Mean: 80.18863049	Standard Deviation: 379.88104639

Attribute hours-per-week
Mean: 38.18863049	Standard Deviation: 11.67713655

Attribute native-country
United-States	Cambodia	England		Puerto-Rico	Canada
0.82227488	0.00236967	0.007109  	0.00947867	0.00473934
Germany		Outlying-US(Guam-USVI-etc)	India		Japan
0.007109  	0.00236967			0.00236967	0.00236967
Greece		South		China		Cuba		Iran
0.00236967	0.00236967	0.00236967	0.007109  	0.00473934
Honduras	Philippines	Italy		Poland		Jamaica
0.00236967	0.007109  	0.00473934	0.00473934	0.00473934
Vietnam		Mexico		Portugal	Ireland		France
0.00236967	0.02843602	0.007109  	0.00236967	0.00236967
Dominican-Republic		Laos		Ecuador		Taiwan
0.00473934			0.00473934	0.00473934	0.00236967
Haiti		Columbia	Hungary		Guatemala	Nicaragua
0.00473934	0.00473934	0.00236967	0.00236967	0.00236967
Scotland	Thailand	Yugoslavia	El-Salvador	Trinadad&Tobago
0.00236967	0.00236967	0.00236967	0.00473934	0.00236967
Peru		Hong		Holand-Netherlands	
0.00236967	0.00236967	0.00236967	


=== Error on training data ===

Correctly Classified Instances         427               85.4    %
Incorrectly Classified Instances        73               14.6    %
Mean absolute error                      0.1624
Root mean squared error                  0.3317
Relative absolute error                 46.3339 %
Root relative squared error             79.3178 %
Total Number of Instances              500     


=== Confusion Matrix ===

   a   b   <-- classified as
  79  34 |   a = >50K
  39 348 |   b = <=50K


=== Error on test data ===

Correctly Classified Instances         415               82.8343 %
Incorrectly Classified Instances        86               17.1657 %
Mean absolute error                      0.1761
Root mean squared error                  0.3559
Relative absolute error                 49.8267 %
Root relative squared error             84.3573 %
Total Number of Instances              501     


=== Confusion Matrix ===

   a   b   <-- classified as
  66  50 |   a = >50K
  36 349 |   b = <=50K

 

by: Keith A. Pray
Last Modified: July 4, 2004 8:59 AM
© 2004 - 1975 Keith A. Pray.
All rights reserved.