k nearest neighbours algorithm python

Question

this is my code for the k nearest neighbor algorithm:

import numpy as np

from EuclideanDistance import EuclideanDistance

dataset = np.loadtxt('C:\Users\Toshiba\Documents\machine learning\RealEstate.csv',      delimiter=',', usecols=(2,3,4,5))
p1 = ()

def normalizeToZscores(data):
    '''Normalizes the variables to z-scores'''
    zScores = list()
for s in data:
    zScore = (s - np.mean(data))/np.std(data)
    zScores.append(zScore) 
    return np.asarray(zScores)

def InOutBudget(data):
    '''Decides whether a particular house is within
    or outside the budget of $153000 and assigns values of 1 and 0 respectively'''
    data2 = list()
    for i in data:
        if (i > 153000): data2.append(0)
        else: data2.append(1) 
    return np.array(data2)

classes = dataset[:,0]
classes = classes.reshape((dataset.shape[0],1))
classes = InOutBudget(classes)

data = dataset[:20,:]
data = normalizeToZscores(data)

p1s = dataset[20:400,:]

def measureDis(data, p1):
    listD = []
    for x in data:
    D = EuclideanDistance(x, p1)
    listD.append(D)
return listD




def most_common(lst):
    '''Finds the most frequently occuring element of a list.
    It will be used to predict a class based on the classification
    of the k-nearest neighbours'''
    return max(set(lst), key=lst.count)

def findKnn(k):
    '''K nearest neighbours algorithm'''
    knns = list()
    errors = list()
    #for i in k:
    for p1 in p1s:
        # Create a list of tuples containing distance and class,
        # Then sort them by shortest distance
        tuples = zip(measureDis(data,p1), classes[20:400])
        tuples = sorted(tuples)
        knn = tuples[:k]
        print knn
        knn = [x[1] for x in knn]
        knn = most_common(knn)
        knns = knns.append(knn)
        print knn
        error = np.abs(knn - p1)
        errors = errors.append(error)
    errorsNum = np.sum(errors)

    return knns

But I keep getting:

Traceback (most recent call last):
File "C:\Users\Toshiba\workspace\assignment5\src\knn2.py", line 76, in <module> knn = findKnn(k)    
File "C:\Users\Toshiba\workspace\assignment5\src\knn2.py", line 64, in findKnn knns = knns.append(knn)
AttributeError: 'NoneType' object has no attribute 'append'

I know the code is really amateur, but could someone please help me just solve the issue?

Are you doing this because you want to implement the algorithm, learn how it works, etc? Or do you need a nearest neighbor search for a larger project? If you just need a nearest neighbor search, take a look at scipy.spatial.KDTree — Elliott
– Elliott, Commented Nov 18, 2013 at 15:59

mdml · Accepted Answer · 2013-11-18 16:00:29Z

6

list.append doesn't return the list. Simply do:

knns.append(knn)

instead of:

knns = knns.append(knn)

edited Nov 18, 2013 at 16:00

mdml

23k8 gold badges61 silver badges66 bronze badges

answered Nov 18, 2013 at 15:59

Exelian

5,9381 gold badge33 silver badges51 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

AShelly · Accepted Answer · 2013-11-18 15:59:37Z

5

append does not return the list, it returns None, so you are clobbering it after the first loop.

answered Nov 18, 2013 at 15:59

AShelly

35.9k17 gold badges98 silver badges163 bronze badges

Collectives™ on Stack Overflow

k nearest neighbours algorithm python

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related