Heart Disease Databases



 

1. Sources:

This directory contains 4 databases concerning heart disease diagnosis:

1. Cleveland Clinic Foundation
2. Hungarian Institute of Cardiology, Budapest
3. V.A. Medical Center, Long Beach, CA
4. University Hospital, Zurich, Switzerland
The authors of the databases are:
  1. Hungarian Institute of Cardiology. Budapest: Andras Janosi, D.
  2. University Hospital, Zurich, Switzerland: William Steinbrunn, M.D.
  3. University Hospital, Basel, Switzerland: Matthias Pfisterer, M.D.
  4. V.A. Medical Center, Long Beach and Cleveland Clinic Foundation.
2. Information Database:             # of instances:

Cleveland:             303
Hungarian:             294
Switzerland:          123
Long Beach VA:   200

Database:             0        1        2        3       4     Total
-----------------------------------------------------------------------
Cleveland:           164     55     36       35     13     303
Hungarian:           188     37     26       28     15     294
Switzerland:         8        48      32       30     5       123
Long Beach VA: 51       56      41       42    10      200
------------------------------------------------------------------------
Total                   411     196    135     135    43

3. Attribute Information:

-- 1. #3 (age)
-- 2. #4 (sex)
-- 3. #9 (cp)
-- 4. #10 (trestbps)
-- 5. #12 (chol)
-- 6. #16 (fbs)
-- 7. #19 (restecg)
-- 8. #32 (thalach)
-- 9. #38 (exang)
-- 10. #40 (oldpeak)
-- 11. #41 (slope)
-- 12. #44 (ca)
-- 13. #51 (thal)
-- 14. #58 (num) (the predicted attribute)

4. Database:

The Heart Database file has 303 + 294 + 123 + 200 = 920 instances. The data required for 100% accuracy is:

Class                 #Required                 #Actual

0                         1400                         411
1                         1400                         196
2                         1400                         135
3                         1400                         135
4                         1400                         43

The first 10 rows are used for the question file. The heart database file has 910 instances. Below are the first 10 instances:

Question                                              Answer

63 1 1 145 233 1 2 150 0 2.3 3 0 6     0
67 1 4 160 286 0 2 108 1 1.5 2 3 3     2
67 1 4 120 229 0 2 129 1 2.6 2 2 7     1
37 1 3 130 250 0 0 187 0 3.5 3 0 3     0
41 0 2 130 204 0 2 172 0 1.4 1 0 3     0
56 1 2 120 236 0 0 178 0 0.8 1 0 3     0
62 0 4 140 268 0 2 160 0 3.6 3 2 3     3
57 0 4 120 354 0 0 163 1 0.6 1 0 3     0
63 1 4 130 254 0 2 147 0 1.4 2 1 7     2
53 1 4 140 203 1 2 155 1 3.1 3 0 7     1

The above 10 rows are used for the question file.

Heart database file.

Question file.
 

5. Results

Click "Average/+ Real" to get the following:

=================== Beginning =====================
 
 

63 1 1 145 233 1 2 150 0 2.3 3 0 6

1.22407

67 1 4 160 286 0 2 108 1 1.5 2 3 3

2.22515

67 1 4 120 229 0 2 129 1 2.6 2 2 7

1.81427

37 1 3 130 250 0 0 187 0 3.5 3 0 3

0.0848638

41 0 2 130 204 0 2 172 0 1.4 1 0 3

0.05125

56 1 2 120 236 0 0 178 0 0.8 1 0 3

0.0515057

62 0 4 140 268 0 2 160 0 3.6 3 2 3

0.0557456

57 0 4 120 354 0 0 163 1 0.6 1 0 3

0.05125

63 1 4 130 254 0 2 147 0 1.4 2 1 7

1.37038

53 1 4 140 203 1 2 155 1 3.1 3 0 7

1.52516

=================== End ==========================

6. Analysis

Correct                 Predicted

0                1.22407 *
2                                            2.22515
1                                            1.81427 *
0                                             0.0848638
0                                             0.05125
0                                             0.0515057
3                                              0.0557456***
0                                             0.05125
2                                             1.37038*
1                                             1.52516*

where:

* miss the target by 1;
*** miss the target by 3.
The result is 5 direct hits, 4 missed by 1, and 1 missed by 3. The number of required instances to achieve 100% accuracy is 1400 per class. The number of actual instances is much smaller than the number of required instances. This in turn limits the DecisionMaker's accuracy. In this case, despite the poverty of the data, the accuracy for class 0 is 80% direct hits and 20% missing the mark by 1.