import math
import csv

class Kneareast:
    def __init__(self, k, lista_treino1, lista_treino2, lista_a_ser_testada1, lista_a_ser_testada2):
        self.k = k
        self.data_base = []
        self.train1 = lista_treino1
        self.test1 = lista_a_ser_testada1
        self.train2 = lista_treino2
        self.test2 = lista_a_ser_testada2


    def management(self):

        for x in self.train1:
            self.data_base.append(x)

        predicts_test1 = self.cicle(self.data_base, self.test1)
        self.sucesso(1, self.train1, predicts_test1)

        predicts_test2 = self.cicle(self.data_base, self.train2)
        self.sucesso(2, self.train2, predicts_test2)

        predicts_test2 = self.cicle(self.data_base, self.test2)
        self.sucesso(3, self.test2, predicts_test2)

        return 0


    def euclidean_distance(self, y1, y2):

        euc = pow(y1 - y2, 2)

        return euc


    def cicle(self, lista1, lista2):

        predicts = []
        i = 0
        while i < len(lista2):

            j = 0
            lista_minimos = []
            while j < len(lista1):

                n = 0
                sumt = 0
                while n < (len(lista1[j])-1):

                    sumt = sumt + self.euclidean_distance(lista1[j][n], lista2[i][n])
                    n = n + 1

                dist = math.sqrt(sumt)

                if len(lista_minimos) >= self.k:
                    if dist < lista_minimos[-1][0]:
                        lista_minimos.pop(-1)
                        lista_minimos.append([dist, lista1[i][-1]])

                else:
                    lista_minimos.append([dist, lista1[i][-1]])

                lista_minimos.sort()
                j = j + 1

            a = 0
            lista_caract = []
            while a < len(lista_minimos):
                lista_caract.append(lista_minimos[a][-1])
                a = a + 1

            max = None
            a = 0
            while a < len(lista_caract):
                if max == None:
                    max = [lista_caract.count(lista_caract[a]), lista_caract[a]]

                else:
                    if max[0] < lista_caract.count(lista_caract[a]):
                        max = [lista_caract.count(lista_caract[a]), lista_caract[a]]
                a = a + 1

            predicts.append(max[1])


            self.data_base.append(lista2[i])

            i = i + 1

        return predicts


    def sucesso(self, number, lista, predicts):

        certo = 0

        i = 0
        while i < len(lista):
            if lista[i][-1] == predicts[i]:
                certo += 1
            i = i + 1

        print(number, predicts, (certo/len(lista))*100)

        return 0

this can be considered as a k-nearest neighbors algorithm in Python?
And why every time I use a base date that consists of 700 elements with this structure: [2.0, 118.0, 80.0, 0.0, 0.0, 42.9, 0.693, 21.0, 1.0], the last digit being the classification of the element, binary classification (1 or 0), no matter how much I increase the number of neighbors there is no difference in the percentage of success, being exactly the same in all attempts? Is some Logic problem?