Esempio n. 1
0
def ada_boosting(args):
    path_train=args.train_file
    path_test=args.test_file
    T=args.iteration
    adaboost=adaboostMM(int(T))
    MNIST_train, Y_train=adaboost.read_MnistFile(path_train)
    adaboost.fit(MNIST_train,Y_train)

    MNIST_test, Y_test=adaboost.read_MnistFile(path_test)
    csoaa_test=adaboost.test_process(MNIST_test)
    pred_label=adaboost.ada_classifier(csoaa_test)
    print 'the accuracy is ', float(sum(pred_label==(Y_test+1)))/len(pred_label)
Esempio n. 2
0
def plot_accuracy_rate(args):
    path_train=args.train_file
    path_test=args.test_file
    accuracys=[]
    T=int(args.iteration)
    for i in range(T):
        adaboost=adaboostMM(T)
        MNIST_train, Y_train=adaboost.read_MnistFile(path_train)
        adaboost.fit(MNIST_train,Y_train)

        MNIST_test, Y_test=adaboost.read_MnistFile(path_test)
        csoaa_test=adaboost.test_process(MNIST_test)
        pred_label=adaboost.ada_classifier(csoaa_test)
        accuracy=float(sum(pred_label==(Y_test+1)))/len(pred_label)
        print accuracy
        accuracys.append(accuracy)
    print accuracys