35 lines
1.0 KiB
JavaScript
35 lines
1.0 KiB
JavaScript
var data = load('test-data-iris.json')
|
|
var t0=time()
|
|
var target = "species";
|
|
var features = ["length", "width","petal_length","petal_width"];
|
|
var datac = ml.preprocess(data,'xy',{features:features,target:target});
|
|
|
|
var model = ml.learn({
|
|
algorithm:ml.ML.ID3,
|
|
data:data,
|
|
target:target,
|
|
features:features
|
|
})
|
|
var t1=time()
|
|
var result = ml.classify(model,datac.x).map(function (r,i) {
|
|
return {value:r,y:datac.y[i]}} )
|
|
var t2=time()
|
|
var correct=0,wrong=0;
|
|
result.forEach(function (r) { if (r.value==r.y) correct++; else wrong++ });
|
|
print('Training Data Test: Correct='+correct+', wrong='+wrong);
|
|
print(toJSON(model).length)
|
|
print(t1-t0,t2-t1)
|
|
datac.x = ml.noise(datac.x,
|
|
{length:0.2,
|
|
width:0.1,
|
|
petal_length:0.05,
|
|
petal_width:0.01
|
|
});
|
|
|
|
// print(datac.x)
|
|
var result = ml.classify(model,datac.x).map(function (r,i) {
|
|
return {value:r,y:datac.y[i]}} )
|
|
correct=0,wrong=0;
|
|
result.forEach(function (r) { if (r.value==r.y) correct++; else wrong++ });
|
|
print('Test Data Test (Noise): Correct='+correct+', wrong='+wrong);
|