dl_tutorial/examples/class_score_predict_sklearn.py at master · mint-lab/dl_tutorial · GitHub

42 lines (35 loc) · 1.2 KB

import numpy as np

import matplotlib.pyplot as plt

from sklearn import svm

midterm_range = np.array([0, 125])

final_range = np.array([0, 100])

# Load score data

class_kr = np.loadtxt('data/class_score_kr.csv', delimiter=',')

class_en = np.loadtxt('data/class_score_en.csv', delimiter=',')

data = np.vstack((class_kr, class_en))

x = data[:,0].reshape(-1, 1)

y = data[:,1]

# Instantiate regression models

models = [

{'name': 'svm.SVR(poly,1)', 'obj': svm.SVR(kernel='poly', degree=1, coef0=1), 'color': 'g'},

{'name': 'svm.SVR(poly,5)', 'obj': svm.SVR(kernel='poly', degree=5, coef0=1), 'color': 'b'},

]

# Prepare for plotting

plt.figure()

plt.plot(data[:,0], data[:,1], 'r.', label='The given data')

idx_sort = np.argsort(data[:,0])

for model in models:

# Train a model

model['obj'].fit(x, y)

# Visualize regression results (the estimated line)

y_pred = model['obj'].predict(x)

model['mae'] = np.mean(np.abs(y_pred - y))

plt.plot(x[idx_sort,0], y_pred[idx_sort], label=f'{model["name"]} (MAE={model["mae"]:.1f})', color=model['color'])

# Decorate the plot

plt.xlabel('Midterm scores')

plt.ylabel('Final scores')

plt.xlim(midterm_range)

plt.ylim(final_range)

plt.grid()

plt.legend()

plt.show()

Expand file tree