train_xgb.py

# prediction=clf.predict(test_features)


import pandas as pd
import numpy as np 
import tensorflow as tf
from sklearn.cross_validation import train_test_split
import matplotlib.pyplot as plt
from sklearn.utils import shuffle
from sklearn.metrics import confusion_matrix
import matplotlib.gridspec as gridspec
from sklearn.preprocessing import StandardScaler
from sklearn.manifold import TSNE
from sklearn.metrics import accuracy_score
from sklearn.metrics import f1_score
from sklearn.metrics import roc_auc_score
import numpy as np 
import pandas as pd
import os
import re
from sklearn.utils import shuffle
import tensorflow as tf
import tflearn
from tflearn.layers.conv import conv_2d, max_pool_2d
from tflearn.layers.core import input_data, dropout, fully_connected
from tflearn.layers.estimator import regression
from tflearn.layers.normalization import local_response_normalization
from sklearn.preprocessing import LabelEncoder
from sklearn.preprocessing import LabelEncoder,OneHotEncoder
import random
import pickle



labelencoder = LabelEncoder() 
onehot_encoder = OneHotEncoder(sparse=False)

VALIDATION_TEST_SET_FILE_LIST = '../validation_feature_batch/audio_7_1080' # ye path change hoga
TRAINING_FILE_LIST = '../train_audio/'  # ye path change hoga
TRAINING_FEATURE_BATCH_PATH = '../train_feature_batch/'  # ye path change hoga
TRAIN_PATH = '/output/audio_classifier/train/'

def one_hot_encode(label_batch):
		label_batch = np.asarray(label_batch)
		integer_encoded = label_batch.reshape(len(label_batch), 1)
		onehot_encoded = onehot_encoder.fit_transform(integer_encoded)
		label_batch = onehot_encoded
		return label_batch



val_feature = [] 
val_labels   = []


val_feature =  pickle.load( open( VALIDATION_TEST_SET_FILE_LIST+ '_feature_batch', "rb" ) )
val_labels =    pickle.load( open(VALIDATION_TEST_SET_FILE_LIST +  '_label_batch', "rb" ) )
val_labels = one_hot_encode(val_labels)


print("training accuracy: "+str(metrics.r2_score(target,pred)))
    
def train_sai_net():

	features = [] 
	target  = []
	prev_best = 0 
	SUBJECT_LIST = [os.path.splitext(subject)[0] for subject in os.listdir(TRAINING_FILE_LIST)]
	# SUBJECT_LIST = ['audio_2_750','audio_3_1368','audio_4_712','audio_5_1020','audio_6_300']
	for subject in SUBJECT_LIST:	
		print(subject)
		feature_batch = pickle.load( open(TRAINING_FEATURE_BATCH_PATH + subject +'_feature_batch' ,  "rb" ) )
		label_batch = pickle.load( open(TRAINING_FEATURE_BATCH_PATH + subject+ '_label_batch', "rb" ) )		  
		label_batch = one_hot_encode(label_batch)
		features.append(feature_batch)
		target.append(label_batch)	
	
	features.append(val_feature)
	label_batch.append(val_labels)


	clf = XGBRegressor(reg_lambda=1,learning_rate=0.03,n_estimators=7500,max_depth=55,min_child_weight=0.7,gamma=0,subsample=0.8,colsample_bytree=0.8, objective= 'binary:logistic' ,scale_pos_weight=1,seed=27)
	clf=clf.fit(features,label_batch,eval_metric='auc')
	pred=clf.predict(features)
	print("training accuracy: "+str(metrics.r2_score(target,pred)))
    label_predict = clf.predict_proba(features)  
    clf.save_model('best_model/0001.model')

train_sai_net()