1

I have an existing model that predicts house prices, that uses simple linear regression. As an input I have date and output is price.

I wanted to improve overall results so I have added one more feature. New feature is distance from the estimated property.

Problem is that the multiple/multivariate regressions performs a bit worse than the simple regression. (All the data are normalised)

Do you have some ideas why is this happening and how can I approach this?

miro
  • 809
  • 10
  • 25

1 Answers1

1

There are dozens of possible reasons, just to list a few:

  • if your new feature is barely correlated with what you are trying to predict - you are efficiently injecting noise to the system thus cannot expect better performance
  • if you have very few data points more features can lead to much harder problem
  • since you are using linear model, even if the new feature is very good predictor, but its relation is not linear to the dependent variable - model will fail as well
  • linear regression as such is very naive model, even ridge/lasso regression might completely change the result (especially lasso since it deals better with bad features)
lejlot
  • 64,777
  • 8
  • 131
  • 164