bistaumanga

#################################################################################
#
# This is FFT library implemented in python.
# The Algorithm used is FFT, decimation in time, radix -2 algorithm
# can compute FFT of 1-d and 2-d lists, 1-d for 1-d signals , and 2-d for images
#
# by : Umanga Bista @ [email protected]
# 
#################################################################################

bistaumanga

'''Implementation and of K Means Clustering
Requires : python 2.7.x, Numpy 1.7.1+'''
import numpy as np

def kMeans(X, K, maxIters = 10, plot_progress = None):

    centroids = X[np.random.choice(np.arange(len(X)), K), :]
    for i in range(maxIters):
        # Cluster Assignment step
        C = np.array([np.argmin([np.dot(x_i-y_k, x_i-y_k) for y_k in centroids]) for x_i in X])

bistaumanga

'''Implementation and  Demostration of Dynamic Time Warping
Requires : python 2.7.x, Numpy 1.7.1, Matplotlib, 1.2.1'''

from math import *
import numpy as np
import sys

def DTW(A, B, window = sys.maxint, d = lambda x,y: abs(x-y)):
    # create the cost matrix
    A, B = np.array(A), np.array(B)

bistaumanga

# -*- coding: utf-8 -*-
"""
Created on Sat May  3 10:21:21 2014

@author: umb
"""

import numpy as np
    
class GMM:

bistaumanga

import numpy as np
from scipy.io import wavfile
from numpy.fft import fft, ifft, fftshift

def hamming(M):
	n = np.arange(M)
	return 0.54 - 0.46 * np.cos( 2 * np.pi * n / (M - 1))

def hann(M):
	n = np.arange(M)

bistaumanga

import numpy as np
import pylab as plt
'''
Performs the Principal Coponent analysis of the Matrix X
Matrix must be n * m dimensions
where n is # features
m is # examples
'''

def PCA(X, varRetained = 0.95, show = False):

bistaumanga

import numpy as np 

def imhist(im):
  # calculates normalized histogram of an image
	m, n = im.shape
	h = [0.0] * 256
	for i in range(m):
		for j in range(n):
			h[im[i, j]]+=1
	return np.array(h)/(m*n)

bistaumanga

# -*- coding: utf-8 -*- 
#############################################################################
#######    usage : > python PR.py -t 0.01 ###################################
####### OR, > python PR.py --thresh=0.01 ####################################
import numpy as np
from optparse import OptionParser

parser = OptionParser()
parser.add_option("-t", "--thresh", dest="threshold", help="Threshold for PR")
options, args = parser.parse_args()

bistaumanga

# -*- coding: utf-8 -*-
"""
Created on Tue Feb 25 22:19:09 2014

@author: bistaumanga

Local Outlier factor implementation in python

Implementations are in two version for calculating knn:
1. Naive method

bistaumanga