Exemple #1
0
def vote(grid_file):
    trees = {}
    x_p, y_p, x_n, y_n = [], [], [], []
    for j in range(1,102):
        trees[outname(j)] = dt.TDIDT(outname(j))
    examples = dt.examplesFromFile(grid_file)
    for true_label in examples:
        for example in examples[true_label]:
            classifications = defaultdict(int)
            for j in range(1,102):
                label = trees[outname(j)].classify(example)
                classifications[label] += 1
            maximum = -sys.maxint-1
            for label in classifications:
                if classifications[label] > maximum:
                    maximum = classifications[label]
                    prediction = label
            if prediction == 1:
                x_p.append(example[0])
                y_p.append(example[1])
            else:
                x_n.append(example[0])
                y_n.append(example[1])
                
    x_p = np.array(x_p)
    y_p = np.array(y_p)
    x_n = np.array(x_n)
    y_n = np.array(y_n)
    
    matlab.plot(x_n,y_n,'go')
    matlab.plot(x_p,y_p,'ro')
    matlab.show() 
Exemple #2
0
def main(data_file="circle.train", grid_file="grid", show=True):
    tdidt = dt.TDIDT(data_file)
    examples = dt.examplesFromFile(grid_file)
    x_p, y_p, x_n, y_n = [], [], [], []
    for label in examples:
        for example in examples[label]:
            prediction = tdidt.classify(example)
            if prediction == 1:
                x_p.append(example[0])
                y_p.append(example[1])
            else:
                x_n.append(example[0])
                y_n.append(example[1])
                
    x_p = np.array(x_p)
    y_p = np.array(y_p)
    x_n = np.array(x_n)
    y_n = np.array(y_n)
    
    matlab.plot(x_n,y_n,'go')
    matlab.plot(x_p,y_p,'ro')
    if show:
        matlab.show()
    else:
        pylab.savefig(data_file, bbox_inches=0)