Add files via upload

PHQ_Patient.csv

@@ -0,0 +1,182 @@
+300_P,2
+301_P,3
+302_P,4
+303_P,0
+304_P,6
+305_P,7
+306_P,0
+307_P,4
+308_P,22
+309_P,15
+310_P,4
+311_P,21
+312_P,2
+313_P,7
+314_P,1
+315_P,2
+316_P,6
+317_P,8
+319_P,13
+320_P,11
+322_P,5
+323_P,1
+324_P,5
+325_P,10
+326_P,2
+327_P,4
+328_P,4
+329_P,1
+330_P,12
+331_P,8
+332_P,18
+333_P,5
+334_P,5
+335_P,12
+336_P,7
+337_P,10
+338_P,15
+339_P,11
+340_P,1
+343_P,9
+344_P,11
+345_P,15
+346_P,23
+347_P,16
+348_P,20
+349_P,5
+350_P,11
+351_P,14
+352_P,10
+353_P,11
+354_P,18
+355_P,10
+356_P,10
+357_P,7
+358_P,7
+359_P,13
+360_P,4
+361_P,0
+363_P,0
+364_P,0
+365_P,12
+366_P,0
+367_P,19
+368_P,7
+369_P,0
+370_P,0
+371_P,9
+372_P,13
+373_P,9
+374_P,2
+375_P,5
+376_P,12
+377_P,16
+378_P,1
+379_P,2
+380_P,10
+381_P,16
+382_P,0
+383_P,7
+384_P,15
+385_P,8
+386_P,11
+387_P,2
+388_P,17
+389_P,14
+390_P,9
+391_P,9
+392_P,1
+393_P,2
+395_P,7
+396_P,5
+397_P,5
+399_P,7
+400_P,7
+401_P,9
+402_P,11
+403_P,0
+404_P,0
+405_P,17
+406_P,2
+407_P,3
+408_P,0
+409_P,10
+410_P,12
+411_P,0
+412_P,12
+413_P,10
+414_P,16
+415_P,3
+416_P,3
+417_P,7
+418_P,10
+419_P,3
+420_P,3
+421_P,10
+422_P,12
+423_P,0
+424_P,3
+425_P,6
+426_P,20
+427_P,5
+428_P,0
+429_P,1
+430_P,3
+431_P,2
+432_P,1
+433_P,10
+434_P,2
+435_P,8
+436_P,0
+437_P,0
+438_P,2
+439_P,1
+440_P,19
+441_P,18
+442_P,6
+443_P,1
+444_P,7
+445_P,1
+446_P,0
+447_P,1
+448_P,18
+449_P,2
+450_P,9
+452_P,1
+453_P,17
+454_P,1
+455_P,1
+456_P,6
+457_P,3
+459_P,16
+461_P,17
+462_P,9
+463_P,0
+464_P,0
+465_P,2
+466_P,9
+467_P,0
+468_P,4
+469_P,3
+470_P,3
+471_P,0
+472_P,3
+473_P,0
+474_P,4
+475_P,6
+476_P,3
+477_P,2
+478_P,1
+479_P,7
+481_P,7
+482_P,1
+483_P,15
+484_P,9
+485_P,4
+486_P,2
+487_P,0
+488_P,0
+489_P,3
+490_P,2
+491_P,8
+492_P,0

Speech duration analysis.py

@@ -0,0 +1,25 @@
+import numpy as np
+import librosa 
+import matplotlib.pyplot as plt
+import os
+import glob
+
+p = []
+q = []
+
+path=  'C:\\Users\\Prajwal\\Desktop\\UMD project\\audio_2\\492_P'
+
+for filename in glob.glob(os.path.join(path, '*.wav')):
+    dur= librosa.get_duration(filename=filename)
+    p.append(filename)
+    q.append(dur)
+
+final=np.empty([len(p), 2])
+array_name=np.asarray(p)
+array_dur=np.asarray(q)
+final=np.concatenate((array_name, array_dur))
+plt.hist(array)
+plt.xlabel('Duration in seconds')
+plt.ylabel('No. of Samples')
+plt.title('Speech duration analysis for P_492')
+plt.savefig('P_492.jpg')

ann_binary_classifier.py

@@ -0,0 +1,112 @@
+import pandas as pd
+import numpy as np
+from sklearn.model_selection import train_test_split
+from sklearn.preprocessing import StandardScaler
+
+l= np.arange(300, 493, 1)
+l=np.delete(l, [18, 21, 41, 42, 62, 94, 98, 151, 158, 160, 180])
+
+path1=[]
+path2=[]
+ffeatures=[]
+foutput=[]
+
+outputs=pd.read_csv('/home/prajwal/UMD_FT/PHQ_Patient.csv', sep=',', header=None)
+outputs=outputs.iloc[:,1]
+output=np.asarray(outputs)
+
+for i in range(len(l)):
+    s1= '/home/prajwal/UMD_FT/transcript/' + str(l[i]) + '_TRANSCRIPT.csv'
+    path1.append(s1)
+    s2= '/home/prajwal/UMD_FT/wwwdaicwoz/' + str(l[i]) + '_P/' + str(l[i]) +'_COVAREP.csv'
+    path2.append(s2)
+
+for j in range(len(path1)): 
+    frames=pd.read_csv(path1[j], sep='\t')
+    features= pd.read_csv(path2[j], sep= ',', header= None)
+    frames1=frames[frames['speaker'].str.match('Participant')]
+    frames2= frames1.iloc[:,0:2]
+    frames3= frames2.values
+    frames4= frames3[:,:]*100
+    c=[] 
+    for i in range(len(frames4)):
+        start_frame= int(frames4[i,0])
+        stop_frame= int(frames4[i,1])
+        a=features.iloc[start_frame:stop_frame, :]
+        c.append(a)
+    final_features=pd.concat(c)
+    ffeatures.append(final_features)
+    arr_ft=np.full([len(final_features), 1], output[j])
+    foutput.append(arr_ft)
+
+ffeatures=pd.concat(ffeatures)
+foutput=np.concatenate(foutput)
+foutput1=foutput.astype(int)
+
+for k in range(len(foutput)):
+   if(foutput[k][0]>10):
+       foutput1[k][0]= 1
+   else:
+       foutput1[k][0]= 0 
+
+ffeatures= StandardScaler().fit_transform(ffeatures)
+ffeatures= pd.DataFrame(data=ffeatures)
+target1=pd.DataFrame(data=foutput1, columns=['Target'])
+finalDf = pd.concat([ffeatures, target1], axis = 1)
+
+depressed= finalDf[(finalDf['Target']==1)]
+control= finalDf[(finalDf['Target']==0)]
+
+depressed1=depressed.sample(n=10000)
+control1= control.sample(n=10000)
+finalDF=pd.concat([depressed1,control1], axis=0)
+finalDF1=finalDF.drop(['Target'], axis=1)
+
+X_train, X_test, y_train, y_test = train_test_split(finalDF1, finalDF['Target'], test_size=0.25)
+
+# Importing the Keras libraries and packages
+from keras.models import Sequential
+from keras.layers import Dense
+
+classifier = Sequential()
+
+# Adding the input layer and the first hidden layer
+classifier.add(Dense(units = 8, kernel_initializer = 'uniform', activation = 'relu', input_dim = 74))
+
+# Adding the second hidden layer
+classifier.add(Dense(units = 16, kernel_initializer = 'uniform', activation = 'relu'))
+classifier.add(Dense(units = 32, kernel_initializer = 'uniform', activation = 'relu'))
+#classifier.add(Dense(units = 64, kernel_initializer = 'uniform', activation = 'relu'))
+#classifier.add(Dense(units = 128, kernel_initializer = 'uniform', activation = 'relu'))
+classifier.add(Dense(units = 16, kernel_initializer = 'uniform', activation = 'relu'))
+# Adding the output layer
+#classifier.add(Dense(units = 32, kernel_initializer = 'uniform', activation = 'relu'))
+classifier.add(Dense(units =8, kernel_initializer = 'uniform', activation = 'relu'))
+classifier.add(Dense(units = 1, kernel_initializer = 'uniform', activation = 'sigmoid'))
+
+# Compiling the ANN
+classifier.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy'])
+
+# Fitting the ANN to the Training set
+classifier.fit(X_train, y_train,validation_data=(X_test,y_test), batch_size =32, epochs =50)
+
+# Part 3 - Making predictions and evaluating the model
+
+# Predicting the Test set results
+y_pred = classifier.predict(X_test)
+from sklearn.metrics import mean_absolute_error
+from sklearn.metrics import mean_squared_error
+import math  
+mae=mean_absolute_error(y_test,y_pred)
+rmse= math.sqrt(mean_squared_error(y_test, y_pred))
+print(mae)
+print(rmse)
+
+#confusion matrix
+y_pred= (y_pred>0.5)
+from sklearn.metrics import confusion_matrix 
+cm = confusion_matrix(y_test, y_pred)
+print(cm)
+acc= (cm[0][0]+cm[1][1])/(cm[0][0]+cm[0][1]+cm[1][0]+cm[1][1])
+print('The accuracy obtained =', acc)
+