vn_score_tab = []
vp_score_tab = []
nu_range = np.logspace(-3, 0, 10)
gamma_range = np.logspace(-2, 0, 6)
# for a in a_range:
for gamma in gamma_range:
print(gamma)
for nu in nu_range:
print(nu)
cl = sksvm.OneClassSVM(gamma=gamma, nu=nu)
cl.fit(x_train)
y_predit = cl.predict(x_test)
vp_score = tools.partial_score(y_test, y_predit, 1)
vn_score = tools.partial_score(y_test, y_predit, -1)
vp_score_tab.append(vp_score)
vn_score_tab.append(vn_score)
print1b1 = False
if print1b1:
for g_range in range(len(gamma_range)):
plt.figure()
plt.semilogx(nu_range,
vn_score_tab[g_range * len(nu_range):(g_range + 1) *
len(nu_range)],
"--",
label=r"Vrai négatif : $\gamma$ = " +
str(gamma_range[g_range]),
# k_range = [1]
# k_range = [2, 3, 4, 5, 6]
k_range = [7, 10, 20, 40, 60, ]
a_range = [0.01, 0.02, 0.03, 0.05, 0.07, 0.09, 0.1, 0.2, 0.3, 0.4, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 5.5, 6, 6.5, 7, 7.5, 8, 8.8, 9, 9.5, 10, 15, 30, 45, 60, 75, 90, 100]
# for a in a_range:
for k in k_range:
print(k)
cl = K_NN_PC(k=k)
cl.fit(x_train)
y_predit = cl.predict_multi_alpha(x_test, a_range)
for a_index in range(len(a_range)):
vp_score = tools.partial_score(y_test, y_predit[:, a_index], 1)
vn_score = tools.partial_score(y_test, y_predit[:, a_index], -1)
vp_score_tab.append(vp_score)
vn_score_tab.append(vn_score)
print1b1 = False
if print1b1:
for k_index in range(len(k_range)):
plt.figure()
plt.semilogx(a_range, vn_score_tab[k_index * len(a_range) : (k_index + 1) * len(a_range)], "--", label=r"Vrai négatif : k = " + str(k_range[k_index]), color=tools.colors[k_index])
plt.semilogx(a_range, vp_score_tab[k_index * len(a_range) : (k_index + 1) * len(a_range)], label=r"Vrai positif : k = " + str(k_range[k_index]), color=tools.colors[k_index])
plt.xlabel(r"Valeur de $\alpha$")
plt.ylabel("Précision")
plt.title(r"Évolution de la précision en fonction du nombre de plus proche voisin et du coefficient $\alpha$.", wrap=True)
plt.legend()
def inlier_score(self, data, label): pred_label = self.predict(data) return tools.partial_score(label, pred_label, 1)
# # # print(cl.score(x_train, y_train))
# # # print(cl.score(x_test, y_test))
# # # print(tools.partial_score(y_test, y_predit, 1))
# # print(np.sum(y_test[y_test == y_predit] == 1))
# # print(tools.partial_score(y_test, y_predit, -1))
cl = nearest_neighbours.Nearest_Neighbours(alpha=4)
cl.fit(x_train, y_train)
# # Fonction spécifique au classifieur Nearest_Neighbours
# # cl.visualisation(x_test)
y_predit = cl.predict(x_test)
vp_score.append(tools.partial_score(y_test, y_predit, 1))
vn_score.append(tools.partial_score(y_test, y_predit, -1))
# # print(cl.score(x_test, y_test))
# # Fonction spécifique au classifieur Nearest_Neighbours
# # print(cl.outlier_score(x_test, y_test))
# # # Fonction spécifique au classifieur Nearest_Neighbours
# # print(cl.inlier_score(x_test, y_test))
# fig, ax = plt.subplots()
# plt.title("Évolution de la précision en fonction du ratio d'exemple négatif disponible")
# ax.scatter(vp_score[:nb_essai], vn_score[:nb_essai], color='r', label="Données artificielles type 1", marker="+")
# ax.scatter(vp_score[nb_essai:], vn_score[nb_essai:], color='b', label="Données artificielles type 2", marker="+")
# plt.ylim(0.45, 1.05)