import numpy as np import matplotlib.pyplot as plt from esnlm.readouts import LogisticRegression input_dim, output_dim = 2, 2 x = np.random.rand(300, input_dim - 1) x = np.hstack([x, np.ones((x.shape[0], 1))]) dmodel = LogisticRegression(input_dim, output_dim) pyr = dmodel.py_given_x(x) y = dmodel.sample_y_given_x(x) model = LogisticRegression(input_dim, output_dim) model.fit(x, y, method='Newton-Raphson', nb_iter=20) py = model.py_given_x(x) plt.plot(x[:, 0], pyr[:, 0], 'x', color='black') plt.plot(x[:, 0], pyr[:, 1], 'x', color='black') plt.plot(x[:, 0], py[:, 0], 'xb') plt.plot(x[:, 0], py[:, 1], 'xg') plt.show()
import numpy as np import matplotlib.pyplot as plt from esnlm.readouts import LogisticRegression input_dim, output_dim = 2, 2 x = np.random.rand(300, input_dim-1) x = np.hstack([x, np.ones((x.shape[0], 1))]) dmodel = LogisticRegression(input_dim, output_dim) pyr = dmodel.py_given_x(x) y = dmodel.sample_y_given_x(x) model = LogisticRegression(input_dim, output_dim) model.fit(x,y, method='Newton-Raphson', nb_iter=20) py = model.py_given_x(x) plt.plot( x[:,0], pyr[:,0], 'x', color='black') plt.plot( x[:,0], pyr[:,1], 'x', color='black') plt.plot( x[:,0], py[:,0], 'xb') plt.plot( x[:,0], py[:,1], 'xg') plt.show()
initial_params = np.array(m.params) m.log_likelihood(x, y) method = 'CG' print "... training using", method, ":", m.fit(x, y, method=method) print "" print "Model Ini ->", init_value print "Model Fin ->", m.log_likelihood(x, y) print "Model Rea ->", dm.log_likelihood(x, y) ############## #### PLOT #### ############## from matplotlib.pyplot import plot, get_cmap, title, show py1 = dm.py_given_x(x) py2 = m.py_given_x(x) for i in range(py1.shape[1]): plot(x[:,0], py1[:, i], 'x', color='black') cm = get_cmap('gist_rainbow') NUM_COLORS = py2.shape[1] for i in range(py2.shape[1]): col = cm(1.*i/NUM_COLORS) plot(x[:,0], py2[:, i], 'x', color=col) title("Multivariable Logistic Regression") show()
initial_params = np.array(m.params) m.log_likelihood(x, y) method = 'CG' print "... training using", method, ":", m.fit(x, y, method=method) print "" print "Model Ini ->", init_value print "Model Fin ->", m.log_likelihood(x, y) print "Model Rea ->", dm.log_likelihood(x, y) ############## #### PLOT #### ############## from matplotlib.pyplot import plot, get_cmap, title, show py1 = dm.py_given_x(x) py2 = m.py_given_x(x) for i in range(py1.shape[1]): plot(x[:, 0], py1[:, i], 'x', color='black') cm = get_cmap('gist_rainbow') NUM_COLORS = py2.shape[1] for i in range(py2.shape[1]): col = cm(1. * i / NUM_COLORS) plot(x[:, 0], py2[:, i], 'x', color=col) title("Multivariable Logistic Regression") show()