Icon

1_​Linear_​Regression_​Housing_​Question

Regression Tree - Ames Housing Data

A regression tree model is trained to predict prices of houses in Ames, Iowa, USA. A number of numerical features are included as predictors

URL: The Learner-Predictor construct in KNIME https://www.youtube.com/watch?v=bKrJkdPvpeA
URL: Slides (Introduction to ML Algorithm course) https://www.knime.com/form/material-download-registration

Exercise: Linear Regression In this exercise we will predict the price of houses given a number of features (size, neighborhood, heating...) using a Linear Regression. 1) Read dataset Housing.csv. It contains information about houses sold (only numerical values) as well as the SalePrice. 2) Remove rows of all columns that have missing values 3) Add Partitioning node to File Reader output a) Top port (training) should have 80 % of the rows b) Draw randomly such rows (set the random seed to ‘1234’) 4) Add Linear Regression Learner to the top output port of Partitioning node a) Select the price column to be learned b) Execute the node 5) Add Regression Predictor a) Predict test set (remaining 20% rows) by simply connecting the remaining unconnected output ports 6) Add Numeric Scorer to Regression Predictor Output a) Reference Column: the target column b) Predicted Column: the new column created by the predictor node

Linear Regresion: SalePrice prediction

Check Assumptions

Check for possibleVariables to reduce
Variance Inflation Factor (VIF)
Calculate Residuals
Math Formula
Check for Heteroscedasticity
Scatter Plot
Check for Normalityof Residuals
Histogram
Normalizer
Check Multicollinearity
Linear Correlation
Read Housing dataset
CSV Reader

Nodes

Extensions

Links