kaggle maching learning笔记

使用scikit建模

1. 定义模型类型，指定模型类型的参数
2. 拟合，从提供的数据捕获模式，建模核心
3. 预测
4. 评估，确定模型的预测的精确度

构建模型
1.导入数据集

import pandas as pd

iowa_file_path = '../input/home-data-for-ml-course/train.csv'

home_data = pd.read_csv(iowa_file_path)

2.指定作为预测目标的属性列y
y = home_data.某一属性列

3.创建训练集X

feature_names = ['LotArea','YearBuilt','1stFlrSF','2ndFlrSF',
                  'FullBath','BedroomAbvGr','TotRmsAbvGrd']

X = home_data[feature_names]

4.分割数据集

from sklearn.model_selection import train_test_split

train_X, val_X, train_y, val_y = train_test_split(X, y, random_state=1)

4.训练模型

• 决策树

from sklearn.tree import DecisionTreeRegressor

iowa_model = DecisionTreeRegressor(random_state=1)

iowa_model.fit(train_X, train_y)

• 随机森林

from sklearn.ensemble import RandomForestRegressor

rf_model = RandomForestRegressor(random_state=1)

5.做出预测
predictions = iowa_model.predict(val_X)

6.评估模型

from sklearn.metrics import mean_absolute_error
val_mae = mean_absolute_error

# uncomment following line to see the validation_mae
#print(val_mae)
step_4.check(predictions, val_y)

调整过拟合和欠拟合的方法
调整决策树max_leaf_nodesDecisionTreeRegressor(max_leaf_nodes=数量)

模板

import pandas as pd

#载入数据集
melbourne_file_path = '../input/melbourne-housing-snapshot/melb_data.csv'
melbourne_data = pd.read_csv(melbourne_file_path) 
#过滤缺值的数据
filtered_melbourne_data = melbourne_data.dropna(axis=0)
#选择目标和属性列
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
#定义模型
melbourne_model = DecisionTreeRegressor()
#训练模型
melbourne_model.fit(X, y)

### 
参数设置
DecisionTreeRegressor(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=False, random_state=None, splitter='best')

###