# use regression X_v, X_a = regression(train_feat, val_mean, aro_mean) # calculating features for whole dataset #print all_feat.shape # use regresion function to calculate v and a all_val = np.sum(np.array(test_feat) * X_v, axis=1) all_aro = np.sum(test_feat * X_a, axis=1) #print all_val.shape #print all_aro.shape print "ATTEMPT" + str(i) avg = average_distance_va(all_val, all_aro, valence, arousal, all_ids) nearest = nearest_dist_average_va(all_val, all_aro, valence, arousal, all_ids) standdev = no_stdev_average_va(all_val, all_aro, val_mean, aro_mean, valence, arousal, all_ids) valence_dist = valence_distance_va(all_val, all_aro, valence, arousal, all_ids) arousal_dist = arousal_distance_va(all_val, all_aro, valence, arousal, all_ids) print 'Average distance: ' + str(avg) print 'Nearest distance: ' + str(nearest) print 'Nearest distance: ' + str(standdev) if avg < best_avg:
# print "check 5" # print X.shape # print len(Yv) # print len(Ya) clf_1 = DecisionTreeRegressor(max_depth=j) clf_2 = DecisionTreeRegressor(max_depth=j) clf_1.fit(Xtrain, Yvtrain) clf_2.fit(Xtrain, Yatrain) Yvpred = clf_1.predict(Xtest) Yapred = clf_2.predict(Xtest) # print len(Yvpred) # print len(Yvtest) # print Yvpred.shape # print Yvtest.shape # avg = averagedist(Yvpred, Yapred, Yvtest, Yatest) avg = average_distance_va(Yvpred, Yapred, valence, arousal, idstest) if avg < best_avg: best_avg = avg print 'j ' + str(j) #print "BEST" print 'Average distance: ' + str(best_avg) # print 'Nearest distance: ' + str(best_near) # print 'Std distance: ' + str(best_std) # print best_val # print best_aro print 'BEST' print 'Average distance: ' + str(best_avg)
# use regression X_v, X_a = regression(train_feat, val_mean, aro_mean) # calculating features for whole dataset #print all_feat.shape # use regresion function to calculate v and a all_val = np.sum(np.array(test_feat) * X_v, axis=1) all_aro = np.sum(test_feat * X_a, axis=1) #print all_val.shape #print all_aro.shape print "ATTEMPT" + str(i) avg = average_distance_va(all_val, all_aro, valence, arousal, all_ids) nearest = nearest_dist_average_va(all_val, all_aro, valence, arousal, all_ids) standdev = no_stdev_average_va(all_val, all_aro, val_mean, aro_mean, valence, arousal, all_ids) valence_dist = valence_distance_va(all_val, all_aro, valence, arousal, all_ids) arousal_dist = arousal_distance_va(all_val, all_aro, valence, arousal, all_ids) print 'Average distance: ' + str(avg) print 'Nearest distance: ' + str(nearest) print 'Nearest distance: ' + str(standdev) if avg < best_avg: best_avg = avg best_near = nearest best_std = standdev
print "check 5" print X.shape print len(Yv) print len(Ya) clf_1 = DecisionTreeRegressor(max_depth=2) clf_2 = DecisionTreeRegressor(max_depth=2) clf_1.fit(Xtrain, Yvtrain) clf_2.fit(Xtrain, Yatrain) Yvpred = clf_1.predict(Xtest) Yapred = clf_2.predict(Xtest) # print len(Yvpred) # print len(Yvtest) # print Yvpred.shape # print Yvtest.shape # avg = averagedist(Yvpred, Yapred, Yvtest, Yatest) avg = average_distance_va(Yvpred, Yapred, valence, arousal, idstest) if avg < best_avg: best_avg = avg print "BEST" print 'Average distance: ' + str(best_avg) # print 'Nearest distance: ' + str(best_near) # print 'Std distance: ' + str(best_std) # print best_val # print best_aro
ids, feat = read_fake_chroma('features/spectrum') train_ids = ids random.shuffle(train_ids) all_ids = train_ids[141:] train_ids = train_ids[0:140] # calcultae valence and arousal find_a_v_mens val_mean, aro_mean = find_a_v_mens_va(train_ids, valence, arousal) train_feat = find_in_dict(feat, train_ids) test_feat = find_in_dict(feat, all_ids) # use regression X_v, X_a = regression(train_feat, val_mean, aro_mean) # calculating features for whole dataset #print all_feat.shape # use regresion function to calculate v and a all_val = np.sum(np.array(test_feat) * X_v, axis=1) all_aro = np.sum(test_feat * X_a, axis=1) #print all_val.shape #print all_aro.shape #plot_all_va(all_val, all_aro, all_ids, valence, arousal) print 'Average distance: ' + str(average_distance_va(all_val, all_aro, valence, arousal, all_ids)) print 'Nearest distance: ' + str(nearest_dist_average_va(all_val, all_aro, valence, arousal, all_ids)) print 'Std distance: ' + str(no_stdev_average_va(all_val, all_aro, val_mean, aro_mean, valence, arousal, all_ids))