learner.stacked: Stacked learner

In MarcoNiemann/kaggle_house: Analysing House Prices in Ames, Iowa and Build a Sales Price Prediction Model

Description

Combines the parameter sets from the tuning of the basic learners to predict the final Saleprice using linear regression.

Usage

learner.stacked(input_train = data_train_numeric_clean_imputed,
  input_test = data_test_numeric_clean_imputed)

Arguments

input_train	Input data which is by default set to data_train_numeric_clean_imputed
input_test	Input data which is by default set to data_test_numeric_clean_imputed

Details

Uses the features_boruta to select only features that are considered important by feature.boruta(). This stacked learner uses xgboost, lasso with the tuned parameters and deeplearning with 10 hidden layers each containing 300 nodes as basic learners. At last linear regression is used to predict the Saleprice using all three predictions. The result is stored in final_submission_stacked_learner.csv and can be directly uploaded to kaggle.

Examples

 KaggleHouse:::learner.stacked()

MarcoNiemann/kaggle_house documentation built on May 7, 2019, 2:50 p.m.