Question

我正在尝试将k最近邻居的代码从R转换为Python。我在最后几行遇到麻烦，分配分类。我对算法有很高的理解，但不确定一些细节。这就是我到目前为止所做的：

from pandas import Series, DataFrame
import pandas as pd
import numpy as np
import numpy.random as npr
from scipy.spatial.distance import pdist, squareform

#define function
def kNN(X,y,k):
    #number of observations
    n = len(y)
    #Set up return values
    y_star = Series(np.zeros(n),dtype='Int64')
    #distance vector
    dist = squareform(pdist(X,metric='Euclidean'))
    #so it can't choose itself as a closest neighbour
    np.fill_diagonal(dist,1e10)
    #Loop through each observation
    for i in range(n):
    #Find the y values of the k nearest neighbours
    y_nearest = y[dist[i,:].argsort()[:k]]
    #assign y_hat
    if y_nearest.mean()> 0.5:
        y_hat[i] = 1
    #return the predicted classification
    return y_hat

测试时给出错误：

ValueError: No axis named 1 for object type <class      'pandas.core.series.Series'>

Python k-最近的邻居

0 个答案: