grohith327

from sklearn.utils import shuffle
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import OneHotEncoder

enc = OneHotEncoder()
y_data = enc.fit_transform(y_data).toarray()
x_data, y_data = shuffle(x_data, y_data)
X_train, X_test, y_train, y_test = train_test_split(x_data, y_data, test_size=0.2,stratify=y_data,random_state=42)

X_train = X_train.reshape((-1,1025,87,1))

grohith327

import librosa
import matplotlib.pyplot as plt 
import numpy as np
import librosa.display

data, sr = librosa.load('Dataset/decrease_volume/1.wav')
D = np.abs(librosa.stft(data))
plt.figure(figsize=(20,8))
librosa.display.specshow(librosa.amplitude_to_db(D,ref=np.max),y_axis='log', x_axis='time')
plt.title('Spectrogram')

grohith327

import numpy as np 
import librosa
import os 


peak_features = []
peak_labels = []
count = 0

for dirc in sorted(os.listdir('./Dataset')):

grohith327

import os 
import librosa
import random

for files in os.listdir('./Dataset'):
    count = 21
    for comm in os.listdir('./Dataset/'+str(files)):
        sound = AudioSegment.from_file('./Dataset/'+str(files)+'/'+str(comm))
        halfway_point = len(sound) // 2 
        for _ in range(10):

grohith327

import pyaudio
import wave
import warnings


warnings.filterwarnings(action='ignore',category=FutureWarning)

CHUNK = 256
FORMAT = pyaudio.paInt16
CHANNELS = 2

grohith327

from google.colab import drive
drive.mount('/content/drive')

grohith327

## K Nearest Neighbors 

y_pred_knn = []
## Iterate through each value in test data 
for val in x_test:
    euc_dis = []
    ## Finding eucledian distance for all points in training data 
    for point in x_train:
        euc_dis.append(((val[0]-point[0])**2+(val[1]-point[1])**2)**0.5)
    temp_target = y_train.tolist()

grohith327

from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import accuracy_score

## We take a range of values for K(1 to 20) and find the accuracy 
## so that we can visualize how accuracy changes based on value of K
accuracy = []
for n in range(1,21):
    clf = KNeighborsClassifier(n_neighbors = n)
    clf.fit(x_train,y_train)
    y_pred = clf.predict(x_test)

grohith327

import numpy as np 
from sklearn.cross_validation import train_test_split

## Retrieve features
X = df.values.tolist()
Y = []
## Convert classes in Strings to Integers 
for val in target:
    if(val == 'Iris-setosa'):
        Y.append(0)

grohith327

import matplotlib.pyplot as plt 

points_1 = df[0:50].values.tolist()
points_2 = df[50:100].values.tolist()
points_3 = df[100:].values.tolist()

points_1 = np.array(points_1)
points_2 = np.array(points_2)
points_3 = np.array(points_3)