kaggle房價預測參考danB

連結：https://www.kaggle.com/learn/machine-learning

以下是用到的輸出

#print(original_data.isnull().sum()) #統計na的總數
#print(original_data.describe())  #顯示描述
#print(original_data.columns) #顯示列
#print(original_data.isnull())   #bool顯示是否為na
#data_without_missing_values = original_data.dropna(axis=1) 刪除na
#print(melbourne_price_data.head())  #預設是前五行
#print(mean_absolute_error(y,predicted_house_prices)) #差值計算
 

下面用隨機森林模型預測，隨機選擇了四個特徵來做，決策樹同隨機森林用法

import pandas as pd
from sklearn.ensemble import RandomForestRegressor

train = pd.read_csv( 'D:/NOTEBOOK/train.csv') #讀取train資料
train_y = train.SalePrice
predictor_x = ['LotArea','YearBuilt','OverallQual','1stFlrSF','FullBath'] #特徵
train_x = train[predictor_x]
my_model = RandomForestRegressor() #隨機森林模型

my_model.fit(train_x,train_y) #fit
test = pd.read_csv( 'D:/NOTEBOOK/test.csv') #讀取test資料
test_x = test[predictor_x]
pre_test_y = my_model.predict(test_x)
print(pre_test_y)

my_submission = pd.DataFrame({'Id':test.Id, 'SalePrice':pre_test_y}) #建csv
my_submission.to_csv('submission2.csv', index=False)