Closed Marslanali closed 4 years ago
Hi,
Thanks for this great library. I have test decision tree C code by porter with Iris data sets, as given in the example and its giving me integrity score of 1. https://github.com/nok/sklearn-porter/tree/stable/examples/estimator/classifier/DecisionTreeClassifier/c
However, when I train the model with my own test data and check the integrity score, its give me zero. And also I run the c code by porter for my custom test data, and its make incorrect predictions.
Below is the my code snippet.
int predict(float features[3]) {
int classes[4]; if (features[1] <= 1738.4554443359375) { if (features[0] <= 88.28286743164062) { if (features[0] <= 84.56035614013672) { classes[0] = 0; classes[1] = 3556; classes[2] = 2; classes[3] = 0; } else { classes[0] = 0; classes[1] = 262; classes[2] = 56; classes[3] = 0; } } else { if (features[2] <= 1944.3283081054688) { classes[0] = 5; classes[1] = 1783; classes[2] = 817; classes[3] = 99; } else { classes[0] = 24; classes[1] = 287; classes[2] = 2300; classes[3] = 305; } } } else { if (features[0] <= 672.5470275878906) { if (features[0] <= 216.8995819091797) { classes[0] = 0; classes[1] = 152; classes[2] = 281; classes[3] = 0; } else { classes[0] = 94; classes[1] = 198; classes[2] = 2; classes[3] = 1992; } } else { if (features[0] <= 4341.470458984375) { classes[0] = 1629; classes[1] = 148; classes[2] = 238; classes[3] = 315; } else { classes[0] = 551; classes[1] = 0; classes[2] = 415; classes[3] = 1198; } } } int index = 0; for (int i = 0; i < 4; i++) { index = classes[i] > classes[index] ? i : index; } return index;
}
Any one please shed some light on this issue ?
Thanks alot
.
Hi,
Thanks for this great library. I have test decision tree C code by porter with Iris data sets, as given in the example and its giving me integrity score of 1. https://github.com/nok/sklearn-porter/tree/stable/examples/estimator/classifier/DecisionTreeClassifier/c
However, when I train the model with my own test data and check the integrity score, its give me zero. And also I run the c code by porter for my custom test data, and its make incorrect predictions.
Below is the my code snippet.
int predict(float features[3]) {
}
Any one please shed some light on this issue ?
Thanks alot