本文數(shù)據(jù)集來自Kaggle波士頓房價(jià)預(yù)測項(xiàng)目https://www.kaggle.com/c/house-prices-advanced-regression-techniques/data
1、數(shù)據(jù)導(dǎo)入

import pandas as pd import matplotlib.pyplot as plt import seaborn as sns import numpy as np from scipy.stats import norm from sklearn.preprocessing import StandardScaler from scipy import stats import warnings warnings.filterwarnings('ignore') %matplotlib inlinetrain = pd.read_csv('input/train.csv') test = pd.read_csv('input/test.csv')

2、數(shù)據(jù)相關(guān)性分析

sample = pd.read_csv('input/sample_submission.csv')corrmat = train.corr() f, ax = plt.subplots(figsize=(12, 9)) sns.heatmap(corrmat, vmax=.8, square=True)

3、挑選相關(guān)變量

k = 10 cols = corrmat.nlargest(k, 'SalePrice')['SalePrice'].index cm = np.corrcoef(train[cols].values.T) sns.set(font_scale=1.25) hm = sns.heatmap(cm, cbar=True, annot=True, square=True, fmt='.2f', annot_kws={'size': 10}, yticklabels=cols.values, xticklabels=cols.values) plt.show()

sns.set() cols = ['SalePrice', 'OverallQual', 'GrLivArea', 'GarageCars', 'TotalBsmtSF', 'FullBath', 'YearBuilt'] sns.pairplot(train[cols], size = 2.5) plt.show()

train_Id = train['Id'] test_Id = test['Id'] len_train = train.shape[0] len_test = test.shape[0] housing = pd.concat([train, test], sort=False) housing.shape (2919, 81) housing[cols].columns Index(['SalePrice', 'OverallQual', 'GrLivArea', 'GarageCars', 'TotalBsmtSF','FullBath', 'YearBuilt'],dtype='object') housing = housing[cols]

4、處理缺失值

housing.isnull().sum() SalePrice 1459 OverallQual 0 GrLivArea 0 GarageCars 1 TotalBsmtSF 1 FullBath 0 YearBuilt 0 dtype: int64 housing.GarageCars = housing.GarageCars.fillna(housing.GarageCars.mean()) housing.TotalBsmtSF = housing.TotalBsmtSF.fillna(housing.TotalBsmtSF.mean()) train = housing[:len_train] test = housing[len_train:]

5、訓(xùn)練模型

xtrain=train.drop("SalePrice",axis=1) ytrain=train['SalePrice'] xtest=test.drop("SalePrice", axis=1) from math import sqrt train_X, val_X, train_y, val_y = train_test_split(xtrain, ytrain, random_state=1) my_pipeline = XGBRegressor(n_estimators=1000, learning_rate=0.05) my_pipeline.fit(train_X, train_y) val_preds = my_pipeline.predict(val_X) msel = mean_squared_error(np.log(val_preds), np.log(val_y)) print("RMSE: %2f" %sqrt(msel)) RMSE: 0.173459 test_preds = my_pipeline.predict(xtest)output = pd.DataFrame({'Id': test_Id,'SalePrice': test_preds}) output.set_index('Id').to_csv('submission.csv')

6、異常值剔除

var = 'GrLivArea' data = pd.concat([train['SalePrice'], train[var]], axis=1) data.plot.scatter(x=var, y='SalePrice', ylim=(0,800000)); 'c' argument looks like a single numeric RGB or RGBA sequence, which should be avoided as value-mapping will have precedence in case its length matches with 'x' & 'y'. Please use a 2-D array with a single row if you really want to specify the same RGB or RGBA value for all points.

train = train.drop(train[train.index == 1298].index) train = train.drop(train[train.index == 523].index) var = 'GrLivArea' data = pd.concat([train['SalePrice'], train[var]], axis=1) data.plot.scatter(x=var, y='SalePrice', ylim=(0,800000)); 'c' argument looks like a single numeric RGB or RGBA sequence, which should be avoided as value-mapping will have precedence in case its length matches with 'x' & 'y'. Please use a 2-D array with a single row if you really want to specify the same RGB or RGBA value for all points.

xtrain=train.drop("SalePrice",axis=1) ytrain=train['SalePrice'] xtest=test.drop("SalePrice", axis=1) from math import sqrt train_X, val_X, train_y, val_y = train_test_split(xtrain, ytrain, random_state=1)my_pipeline = XGBRegressor(n_estimators=1000, learning_rate=0.05) my_pipeline.fit(train_X, train_y)val_preds = my_pipeline.predict(val_X) msel = mean_squared_error(np.log(val_preds), np.log(val_y)) print("RMSE: %2f" %sqrt(msel)) RMSE: 0.162534 test_preds = my_pipeline.predict(xtest)output = pd.DataFrame({'Id': test_Id,'SalePrice': test_preds}) output.set_index('Id').to_csv('submission2.csv')

最終結(jié)果0.169，排名75%。后續(xù)優(yōu)化方向，變量選擇、特征工程、模型選擇。

總結(jié)

以上是生活随笔為你收集整理的kaggle房价预测特征意思_Kaggle实战-波士顿房价预测的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯(cuò)，歡迎將生活随笔推薦給好友。