Day 4 [Python ML] 模型验证

前言

今天要继续昨天做过的部分，因此一开始需要昨天的程序码

import pandas as pd

# 读起资料
melbourne_file_path = './Dataset/melb_data.csv'
melbourne_data = pd.read_csv(melbourne_file_path) 
# drop缺失值
filtered_melbourne_data = melbourne_data.dropna(axis=0)
# 选择target和feature
y = filtered_melbourne_data.Price
melbourne_features = ['Rooms', 'Bathroom', 'Landsize', 'BuildingArea', 
                        'YearBuilt', 'Lattitude', 'Longtitude']
X = filtered_melbourne_data[melbourne_features]

from sklearn.tree import DecisionTreeRegressor
# 定义model
melbourne_model = DecisionTreeRegressor()
# Fit这个模型
melbourne_model.fit(X, y)

DecisionTreeRegressor(ccp_alpha=0.0, criterion='mse', max_depth=None,
                      max_features=None, max_leaf_nodes=None,
                      min_impurity_decrease=0.0, min_impurity_split=None,
                      min_samples_leaf=1, min_samples_split=2,
                      min_weight_fraction_leaf=0.0, presort='deprecated',
                      random_state=None, splitter='best')

平均绝对误差

若资料为数值属性，则可以取得每笔资料的误差，全部相加後做平均，即可得到评估Model好坏的标准

from sklearn.metrics import mean_absolute_error

predicted_home_prices = melbourne_model.predict(X)
mean_absolute_error(y, predicted_home_prices)

434.71594577146544

切分train和test资料

若是验证时使用的资料集和训练时一样，称为in-sample

这种情况会造成计算出来的结果非常的好，但却不贴近真实情况

因此需要将资料分为训练资料以及测试资料

from sklearn.model_selection import train_test_split

# 切分资料成训练资料和验证资料，feature和target都要切分
# 这是使用随机切分的，设定random_stated可以确保每次切分的资料都是一样的 
train_X, val_X, train_y, val_y = train_test_split(X, y, random_state = 0)
# 定义model
melbourne_model = DecisionTreeRegressor()
# Fit这个模型
melbourne_model.fit(train_X, train_y)

# 得到验证资料的结果
val_predictions = melbourne_model.predict(val_X)
print(mean_absolute_error(val_y, val_predictions))