【分類算法】KNN

import numpy as np
import pandas as pd
from sklearn.model_selection import train_test_split
from scipy.spatial.distance import cdist
class KNN(object):
    def __init__(self, k=5):
        self.k = k
        self.DataSet = list()
        self.Distance = list()
        self.TableNN = list()
        self.Result = list()
        pass

    def fitting(self, X, Y):
        """ 建模 """
        self.Distance = cdist(X, X, metric='euclidean')
        for Index in range(len(X)):
            self.DataSet.append({'Index': Index,
                                 'Data': X[Index],
                                 'Signal': Y[Index]})
            self.TableNN.append([])
            for nn in sorted(enumerate(self.Distance[Index]), key=lambda X: X[1])[1:self.k + 1]:
                self.TableNN[-1].append({'Index': Index, 'Target': nn[0], 'Distance': nn[1]})

        self.TableNN = np.array(self.TableNN)
        pass

    def Forecast(self, X):
        DISTANCE = cdist(np.array(X), np.array(pd.DataFrame(self.DataSet)['Data'].tolist()), metric='euclidean')
        for dist in DISTANCE:
            """ 計算距離 """
            Type = dict()
            for nn in sorted(enumerate(dist), key=lambda X: X[1])[1:self.k + 1]:
                """ 通過序號查找索引，進而獲取近鄰類別 """
                if self.DataSet[nn[0]]['Signal'] in Type.keys():
                    Type[self.DataSet[nn[0]]['Signal']] = Type[self.DataSet[nn[0]]['Signal']] + 1
                else:
                    Type[self.DataSet[nn[0]]['Signal']] = 1
            self.Result.append(sorted(Type.items(), key=lambda x: x[0])[0][0])

        return self.Result

    def accuracy(self, Y):
        trueNum = 0
        for Index in range(len(self.Result)):
            if self.Result[Index] == Y[Index]:
                trueNum = trueNum + 1

        print("The accuracy of KNN is: {}".format(trueNum / len(self.Result)))
        pass