Induction Engine Demo: Train File Report

Return to Conclusion.
 
Sample Size = 250
Decision Tree (Entropy):

    0.790323 0.202555


Sample Size = 250
Decision Tree (C4.5):

    0.943548 0.0285937


Sample Size = 250
Naive Models Algorithm:

    0.834677 0.0437485


Sample Size = 250
Decision Tree (Gini):

    0.802419 0.204336


Sample Size = 250
Naive Bayesian Algorithm:

    0.858871 0.0595746


Sample Size = 250
Logistic Regression Algorithm:

Threshold 0.3 0.879032 0.0591049
Threshold 0.5 0.879032 0.0480015
Threshold 0.7 0.875 0.0608094


Sample Size = 250
k-Nearest Neighbor Algorithm:

1   0.790323 0.0808752
3   0.814516 0.0615686
5   0.826613 0.0515481
7   0.850806 0.0485787
10   0.83871 0.0344853
15   0.834677 0.0437485
20   0.834677 0.0437485
25   0.834677 0.0437485



Sample Size = 500
Decision Tree (Entropy):

    0.766129 0.11821


Sample Size = 500
Decision Tree (Gini):

    0.707661 0.163905


Sample Size = 500
Decision Tree (C4.5):

    0.933468 0.0338737


Sample Size = 500
Naive Bayesian Algorithm:

    0.816532 0.0683957


Sample Size = 500
Naive Models Algorithm:

    0.802419 0.0750353


Sample Size = 500
Logistic Regression Algorithm:

Threshold 0.3 0.858871 0.0307843
Threshold 0.5 0.868952 0.0327582
Threshold 0.7 0.828629 0.0479531


Sample Size = 500
k-Nearest Neighbor Algorithm:

1   0.820565 0.0451592
3   0.828629 0.0479531
5   0.830645 0.0645161
7   0.826613 0.0632068
10   0.824597 0.0618321
15   0.816532 0.082214
20   0.808468 0.0773221
25   0.802419 0.0750353



Sample Size = 1000
Naive Models Algorithm:

    0.795 0.0622529


Sample Size = 1000
Decision Tree (C4.5):

    0.943 0.029761


Sample Size = 1000
Naive Bayesian Algorithm:

    0.841 0.0397564


Sample Size = 1000
Decision Tree (Gini):

    0.82 0.0624088


Sample Size = 1000
Decision Tree (Entropy):

    0.79 0.0565281


Sample Size = 1000
Logistic Regression Algorithm:

Threshold 0.3 0.889 0.0265115
Threshold 0.5 0.887 0.0215672
Threshold 0.7 0.845 0.0382324


Sample Size = 1000
k-Nearest Neighbor Algorithm:

1   0.834 0.0429219
3   0.849 0.0337893
5   0.831 0.0309654
7   0.83 0.0437656
10   0.837 0.0336537
15   0.828 0.0442331
20   0.824 0.0507768
25   0.818 0.0565281



Sample Size = 2000
Decision Tree (C4.5):

    0.9545 0.0207227


Sample Size = 2000
Decision Tree (Entropy):

    0.794 0.0467455


Sample Size = 2000
Naive Models Algorithm:

    0.786 0.0329848


Sample Size = 2000
Decision Tree (Gini):

    0.799 0.0735002


Sample Size = 2000
Naive Bayesian Algorithm:

    0.8255 0.0222133


Sample Size = 2000
Logistic Regression Algorithm:

Threshold 0.3 0.86 nan
Threshold 0.5 0.892 nan
Threshold 0.7 0.888 nan


Sample Size = 2000
k-Nearest Neighbor Algorithm:

1   0.8405 0.0135962
3   0.856 0.0294715
5   0.8565 0.0279131
7   0.8515 0.0205565
10   0.849 0.0200286
15   0.8345 0.0219024
20   0.831 0.0250941
25   0.8175 0.0294667



Sample Size = 4000
Naive Bayesian Algorithm:

    0.84925 0.0410287


Sample Size = 4000
Decision Tree (Entropy):

    0.82175 0.0496552


Sample Size = 4000
Decision Tree (Gini):

    0.83975 0.0440187


Sample Size = 4000
Naive Models Algorithm:

    0.8015 0.0197629


Sample Size = 4000
Decision Tree (C4.5):

    0.965 0.00748331


Sample Size = 4000
k-Nearest Neighbor Algorithm:

1   0.86 0.0148131
3   0.87125 0.0107935
5   0.86525 0.0131774
7   0.86525 0.0127811
10   0.8675 0.0178165
15   0.8565 0.0166562
20   0.856 0.0183615
25   0.84675 0.0187598



Return to Conclusion.


©2002 PredictionWorks, Inc. All rights reserved.