Day 02. KNN Practice with Spine Dataset

# -*- coding: utf-8 -*-
"""SpineDataset.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1VLXQO-yB6QUICDdMF0ur49Du-gFAlAAL
"""
 
from google.colab import drive
drive.mount('/gdrive')
 
# /gdrive/My Drive/Colab Notebooks/resources/ <==  My resource path
 
import numpy as np
import matplotlib.pyplot as plt
import time
 
 
labels = [ b'NO', b'DH', b'SL' ]
data = np.loadtxt("/gdrive/My Drive/Colab Notebooks/resources/SPINE/column_3C.dat", converters={6: lambda s: labels.index(s)} )
 
data.shape
 
x = data[: , 0:6]
y = data[: ,   6]
 
training_indices = list(range(0 , 20)) + list(range(  4,188)) + list(range(230, 310))
test_indices     = list(range(20, 40)) + list(range(188,230))
 
trainx = x[training_indices,:]
trainy = y[training_indices  ]
testx  = x[test_indices,:    ]
testy  = y[test_indices      ]
 
"""##Functions"""
 
def NN_L2(trainx, trainy, testx):
    
    predicY = np.array([ trainy[ find_nn(testI) ] for testI in testx  ]) 
        
    return predicY
 
def find_nn(x):
    distances = [ l2_dist(x, trainx[i, ]) for i in range(len(trainx)) ]
    
    return np.argmin(distances)
 
def l2_dist(v1, v2):
    return np.sum( np.square(v1-v2) )
 
def getLabel(idx):
    return trainy[idx]
 
def NN_L1(trainx, trainy, testx):
    
    predicY = np.array([ trainy[ find_nnL1(testI) ] for testI in testx  ]) 
        
    return predicY
 
def find_nnL1(x):
    distances = [ l1_dist(x, trainx[i, ]) for i in range(len(trainx)) ]
    
    return np.argmin(distances)
 
def l1_dist(v1, v2):
    return np.sum( np.abs(v1-v2) )
 
def confusion(predicLabel, testLabel):
    confMat = np.zeros((3,3) , dtype = np.int32)
    
    for i in range(len(testLabel)):
            
        confMat[int(predicLabel[i]), int(testLabel[i])] += 1
    
    return confMat
 
"""##Main"""
 
testy_L2 = NN_L2(trainx, trainy, testx)
print( type( testy_L2) )
print( len(testy_L2) )
print( testy_L2[40:50] )
 
err_positions = np.not_equal(testy_L2, testy)
error         = float(np.sum(err_positions)) / len(testx)
 
error
 
testy_L1 = NN_L1(trainx, trainy, testx)
 
print( type( testy_L1) )
print( len(testy_L1) )
print( testy_L1[40:50] )
print( all(testy_L1 == testy_L2) )
 
err_positions = np.not_equal(testy_L1, testy)
error         = float(np.sum(err_positions)) / len(testx)
 
error
 
resultL2 = confusion(testy_L2, testy)
resultL1 = confusion(testy_L1, testy)
 
print(resultL2)
print(resultL1)
 

출처 및 참고자료 : edx -  Machine Learning Fundamentals_week_1 Programming Assignment.2