def svm_3_pca(train_x,train_y,kaggle_test): pcas = [100,500,1000] gammas = [0.001,0.01,0.1] cs = [1.0,10.0,100.0] accuracy_list = [] m,n = train_x.shape kf = KFold(m, n_folds=2) for k in pcas: print k,'pca' accuracy_list_p = [] train_x_transform,pca_f = pca(train_x,k) kaggle_test_transform = pca_f.transform(kaggle_test) # X_train, X_valid, Y_train, Y_valid = train_test_split(train_x_transform, train_y,test_size=0.2,random_state=0) for g in gammas: for c in cs: accs = [] counter = 0 for train_index, test_index in kf: X_train, X_test,y_train, y_test = train_x[train_index],train_x[test_index],train_y[train_index],train_y[test_index] clf = svm.SVC(C= c,gamma=g) clf.fit(X_train,y_train) pred = clf.predict(X_test) # print type(pred) corr, acc = accuracy(y_test,pred) accs.append(acc) print "results of svm using pca "+str(counter)+'counter '+str(k)+'pca components '+str(c)+'c'+str(g)+'gamma ' + ":\n%s\n" %(metrics.classification_report(y_test,pred)) fname2 = 'results/svm_pca/'+str(counter)+'counter'+str(k)+'pca components'+str(c)+'c'+str(g)+'gamma'+'cfm_svm_pca.png' cfm_svm_pca(pred,y_test,c,g,k,fname2) counter +=1 acc_mean = np.mean(accs) accuracy_list.append((k,c,g,acc_mean)) accuracy_list_p.append((c,g,acc_mean)) fname = 'results/svm_pca/svm_pca_accuracy_'+str(k)+'pca.png' plot_svm_accuracy(accuracy_list_p,fname) sortedaccuracy_list = sorted(accuracy_list, key = lambda x:x[3], reverse = True) print sortedaccuracy_list pca_optimal = sortedaccuracy_list[0][0] c_optimal = sortedaccuracy_list[0][1] g_optimal = sortedaccuracy_list[0][2] train_x_transform,pca_f = pca(train_x,pca_optimal) kaggle_test_transform = pca_f.transform(kaggle_test) print "classifcation kaggle" clf = svm.SVC(C= c_optimal,gamma=g_optimal) clf.fit(train_x_transform,train_y) pred_kaggle_svm = clf.predict(kaggle_test_transform) fname3 = 'results/svm_pca/svm_pca_kaggle.csv' write_tokaggle(fname3,pred_kaggle_svm)
def logistic_regression2_pca(train_x,train_y,kaggle_test): pcas = [100,500,1000] m,n = train_x.shape kf = KFold(m, n_folds=2) cs = [1.0,10.0,100.0] accuracy_list=[] for k in pcas: for c in cs: accs = [] counter = 0 for train_index, test_index in kf: print "counter" + str(counter) X_train, X_test,y_train, y_test = train_x[train_index],train_x[test_index],train_y[train_index],train_y[test_index] clf = linear_model.LogisticRegression(C = c) clf.fit(X_train,y_train) pred = clf.predict(X_test) corr, acc = accuracy(y_test,pred) accs.append(acc) print "results of logistic_regression using pca "+str(counter)+'counter '+str(k)+'pca components '+str(c)+'lambda'+":\n%s\n" %(metrics.classification_report(y_test,pred)) fname2 = 'results/logistic_regression_pca/'+str(counter)+'counter'+str(k)+'pca components'+str(c)+'lambda'+'cfm_logistic_regression_pca.png' cfm_logistic_regression_pca(pred,y_test,c,k,fname2) counter += 1 acc_mean = np.mean(accs) accuracy_list.append((k,c,acc_mean)) sortedaccuracy_list = sorted(accuracy_list, key = lambda x:x[2], reverse = True) print sortedaccuracy_list c_optimal = sortedaccuracy_list[0][1] pca_optimal = sortedaccuracy_list[0][0] fname = 'results/logistic_regression_pca/logistic_regression_accuracy_pca.png' plot_logistic_regression_accuracy_pca(accuracy_list,fname) train_x_transform,pca_f = pca(train_x,pca_optimal) kaggle_test_transform = pca_f.transform(kaggle_test) print "classifcation kaggle" clf = linear_model.LogisticRegression(C= c_optimal) clf.fit(train_x_transform,train_y) pred_kaggle_svm = clf.predict(kaggle_test_transform) fname3 = 'results/logistic_regression_pca/logistic_regression_pca_kaggle.csv' write_tokaggle(fname3,pred_kaggle_svm)
from load_numpy import getNumpy from test_classifier import svm1 from preprocesses import pca from rbm import rbm from sklearn.cross_validation import train_test_split train_x,train_y,test_x = getNumpy() train_x_transform = pca(train_x,100) X_train, X_valid, Y_train, Y_valid = train_test_split(train_x_transform, train_y,test_size=0.2,random_state=0) # svm1(train_x_transform,train_y) rbm(train_x,train_y)