K近邻KNeighborsRegressor--StandardScaler标准化--mean_squared_error均方根误差 学习笔记

本文主要是介绍K近邻KNeighborsRegressor--StandardScaler标准化--mean_squared_error均方根误差 学习笔记，希望对大家解决编程问题提供一定的参考价值，需要的开发者们随着小编来一起学习吧！

np.abs()函数

np.abs() : 计算数值各元素运算的绝对值

import numpy as np
np.abs([1,-2,3,-4])Out:
array([1, 2, 3, 4])

• np.sqrt() : 计算平方根
• np.square() : 计算平方

pd.sample()参数含义

DataFrame.sample(n=None, frac=None, replace=False, weights=None, random_state=None, axis=None)[source]
DataFrame可以是Series、DataFrame

• n的含义是抽样的个数，是整数；frac是浮点数，是抽样的比例
• replace为True含义为数据本身改变，为False含义为数据本身未改变 , 需要定义新变量接收
• weights的含义是给抽样所在axis的每个元素赋值抽样权重，所以weights的长度必须和所在axis的长度相同，不然会报错，缺失值的weights会被设置为0，如果weights加和不等于1，会被normalized到加和为1，inf和-inf值不被允许
• axis的含义是抽样的方向，axis=0，对行进行抽样，axis=1，对列进行抽样
• random_state是用来复现结果的

pd.str同时去掉分隔符和货币符号

DataFrame数据类型有千位分隔符和货币符 , 即为"$10,000.00"时 , 怎么转换为int或者float类型 :

• df.str.replace("$|," , “”).astype(float)

standarscaler注意点

• StandarScaler().fit_transform(a) 中的a不能直接是一个DataFrame , 而要是DataFrame中的具体几列

scipy.spatial中distance距离工具

from scipy.spatial import distance

两点之间的距离

• 使用distance.euclidean(a,b)

两个数据之间的距离

• 使用distance.cdist ( m[a] , n[b] )

使用Sklearn计算距离

from sklearn.neighbors import KNeigborsRegressor

knn = KNeighborsRegressor()
cols = ['A','B']
knn.fit(train_data[cols],train_data['C'])
ret = knn.predict(test_df[cols]

sklearn 计算均方根误差

from sklearn.metrics import mean_squared_error

mse = mean_squared_error( test_data['C'], ret)
rmse = mse ** (1/2)

sklearn标准化

from sklearn.preprocessing import StandarScater
data[features] = StandardScaler().fit_transform(data[features])

features为数据data的columns

K近邻模型

多变量knn模型测试

import pandas as pd
from sklearn.preprocessing import StandardScalerdc_data=pd.read_csv('listings.csv')
features=['accommodates','bedrooms','bathrooms','beds','price','minimum_nights','maximum_nights','number_of_reviews']
dc_data = dc_data[features]
dc_data = dc_data.dropna() #去掉NaN数据
dc_data['price'] = dc_data.price.str.replace('\$|,','').astype(float) #将货币千分位数转换为float类数据
dc_data[features] = StandardScaler().fit_transform(dc_data[features]) #标准化
normalized_data = dc_data# 建立训练集和测试集
norm_train_df = normalized_data.copy().iloc[:2750]
norm_test_df = normalized_data.copy().iloc[2750:]from scipy.spatial import distancedef predict_price_multivariate(new_data_value,feature_columns):temp_df = norm_train_dftemp_df['distance'] = distance.cdist(temp_df[feature_columns],[new_data_value[feature_columns]])temp_df = temp_df.sort_values('distance')knn_5 = temp_df.price.iloc[:5]predicted_price = knn_5.mean()return predicted_price
# K近邻预测价格的函数cols = ['accommodates', 'bathrooms']
norm_test_df['predicted_price'] = norm_test_df[cols].apply(predict_price_multivariate,feature_columns=cols,axis=1)
norm_test_df['squared_error'] = (norm_test_df['predicted_price']-norm_test_df['price'])**2
mse = norm_test_df['squared_error'].mean()
rmse = mse ** (1/2)
# 最终获得RMSE均方根误差值

0.15274960512174854

K近邻

import pandas as pd
from sklearn.preprocessing import StandardScaler
from sklearn.neighbors import KNeighborsRegressorfeatures = ['accommodates','bedrooms','bathrooms','beds','price','minimum_nights','maximum_nights','number_of_reviews']
dc_data = pd.read_csv('listings.csv')
dc_data = dc_data[features]
dc_data['price'] = dc_data.price.str.replace('\$|,','').astype(float)
dc_data = dc_data.dropna()dc_data[features] = StandardScaler().fit_transform(dc_data[features])
normalized_data = dc_data
# 标准化norm_train_df = normalized_data.copy().iloc[:2750]
norm_test_df = normalized_data.copy().iloc[2750:]cols = ['accommodates','bedrooms','bathrooms','beds','minimum_nights','maximum_nights','number_of_reviews']
knn = KNeighborsRegressor(n_neighbors=4) #默认n_neighbors值为5
knn.fit(norm_train_df[cols], norm_train_df['price'])
four_features_predictions = knn.predict(norm_test_df[cols])from sklearn.metrics import mean_squared_error
four_features_mse = mean_squared_error(norm_test_df['price'],four_features_predictions)
four_features_rmse = four_features_mse ** (1/2)
four_features_rmse

0.823713617207827

这篇关于K近邻KNeighborsRegressor--StandardScaler标准化--mean_squared_error均方根误差 学习笔记的文章就介绍到这儿，希望我们推荐的文章对编程师们有所帮助！

---