spotify_knn.py

# Importing libraries
import pandas as pd
import numpy as np
import math
import operator


data = pd.read_csv('training_data.csv')

# Defining a function which calculates euclidean distance between two data points
def euclideanDistance(data1, data2, length): # individually weighting the chosen parameters
    distance = 0
    distance += (1.5*(np.square(data1[0] - data2[0])))  #danceability
    distance += np.square(data1[1] - data2[1])  #acousticness
    distance += (1.5*(np.square(data1[2] - data2[2])))  #energy
    distance += np.square(data1[3] - data2[3])  #instrumentalness
    distance += (0.5*(np.square(data1[4] - data2[4])))  #liveness
    distance += (2*(np.square(data1[5] - data2[5])))  #valence
    distance += (1.5*(np.square(data1[6] - data2[6])))  #loudness
    distance += np.square(data1[7] - data2[7])  #speechiness
    distance += (1.5*(np.square(data1[8] - data2[8])))  #tempo
    return np.sqrt(distance)

# Defining the KNN model
def knn(trainingSet, testInstance, k):
    distances = {}
    sort = {}
    length = testInstance.shape[0]
    
    # Calculating euclidean distance between each row of training data and test data
    for x in range(len(trainingSet)):
        dist = euclideanDistance(testInstance, trainingSet.iloc[x], length)
        distances[x] = dist
 
    # Sorting them on the basis of distance
    sorted_d = sorted(distances.items(), key=operator.itemgetter(1))
 
    neighbors = []
    
    # Extracting top k neighbors
    for x in range(k):
        neighbors.append(sorted_d[x][0])

    classVotes = {}
    
    # Calculating the most freq class in the neighbors
    for x in range(len(neighbors)):
        response = trainingSet.iloc[neighbors[x]][-1]
 
        if response in classVotes:
            classVotes[response] += 1
        else:
            classVotes[response] = 1

    sortedVotes = sorted(classVotes.items(), key=operator.itemgetter(1), reverse=True)
    return(sortedVotes[0][0], neighbors)

# Test set contains data from 'test_data.csv' manually inputted/written here
testSet = [[0.063819, 0.000009, 0.089448, 0.000012, 0.019381, 0.079786, -0.028829, 0.040249, 0.058942], [0.05518, 0.001079, 0.057774, 0, 0.031516, 0.05568, -0.025444, 0.018343, 0.047111], [0.066519, 0.070958, 0.018763, 0.07414, 0.021615, 0.010127, -0.062838, 0.029031, 0.065932], [0.081421, 0.001859, 0.07468, 0.006605, 0.051369, 0.127342, -0.031939, 0.027894, 0.062684], [0.038227, 0.000024, 0.091213, 0.000032, 0.050625, 0.025416, -0.013444, 0.127341, 0.067778], [0.056584, 0.001489, 0.085361, 0.000008, 0.032509, 0.111097, -0.024457, 0.033048, 0.041453], [0.065331, 0.075656, 0.02963, 0.000259, 0.028787, 0.035766, -0.04839, 0.023194, 0.051669], [0.070514, 0.000775, 0.084247, 0.089668, 0.084871, 0.031181, -0.02611, 0.083378, 0.06038], [0.036607, 0.093645, 0.01319, 0.09951, 0.021019, 0.015852, -0.096021, 0.031532, 0.051847], [0.083364, 0.049071, 0.023779, 0.000669, 0.035735, 0.046771, -0.037963, 0.036459, 0.056053], [0.080881, 0.006736, 0.064927, 0.00021, 0.018935, 0.122495, -0.058236, 0.028121, 0.060768], [0.017062, 0.096443, 0.008945, 0.011482, 0.018662, 0.005175, -0.080969, 0.030168, 0.033477], [0.063711, 0.027684, 0.076909, 0.000001, 0.02154, 0.083323, -0.0279, 0.033275, 0.052494], [0.059716, 0.086149, 0.022292, 0.103447, 0.023129, 0.021486, -0.086731, 0.023043, 0.060807], [0.065007, 0.009914, 0.086011, 0.00491, 0.059807, 0.0752, -0.007197, 0.067991, 0.047149], [0.070298, 0.019888, 0.046721, 0.070422, 0.019381, 0.073104, -0.048453, 0.024483, 0.06272], [0.022785, 0.090247, 0.001365, 0.102244, 0.084871, 0.008738, -0.102023, 0.034564, 0.080094], [0.061767, 0.000018, 0.091956, 0.000032, 0.029283, 0.052011, -0.00885, 0.050027, 0.050401], [0.034555, 0.063962, 0.023779, 0.094261, 0.027794, 0.025023, -0.079881, 0.03373, 0.054219], [0.047945, 0.044874, 0.046164, 0.078952, 0.020101, 0.032753, -0.055983, 0.029031, 0.046715]]
for i in range(20):
    test = np.array(testSet[i])

    k = 5 # chosen k value
    result,neigh = knn(data, test, k) 
    print('\nPredicted class of the datapoint (5) = ',result)