1

first post here!

I have a project to estimate commercial real estate properties given categorical and continuous variables. I have ran a step-wise linear regression model to pick out a good formula in RStudio. (Adj. Rsq = .90 also, I know I need to do PCA and some type of categorical ANOVA test still, but I just want to get a beta estimator out before going deeper.)

  • I'd like to figure out how to correctly deal with the interactive variables that have a transformed factor in it. Ex: log(Sqft):floors:building_materialWood from R. Normally it would be a simple natural log to cancel out the transformation but I feel like that is an inappropriate solution for interaction variables.

Thank you so much! If anything is not specific enough I will do my best to further explain.

S_Stand_ring
  • 11
  • 3
  • What do you mean by LM? – Michael R. Chernick Mar 29 '17 at 22:57
  • 1
    LM is the *R* function for Linear Model – S_Stand_ring Mar 29 '17 at 23:13
  • In that case maybe you should migrate the question to another site. I would suggest stackoverflow if you can add replicable R code. – Michael R. Chernick Mar 29 '17 at 23:31
  • @MichaelChernick Ok thanks. I'll definitely try that. – S_Stand_ring Mar 31 '17 at 01:53
  • @MichaelChernick is this better suited for [Cross Validated](http://stats.stackexchange.com)? – S_Stand_ring Mar 31 '17 at 15:43
  • This seems more like a statistics question than a coding question, so it seems better here than on Cross Validated. It would help to see the model you have to this point and some output from it; the explanation in text above isn't completely clear. Also, note that stepwise regression is probably not a good idea, as you can tell from following that tag on the site. Say more about what you are trying to accomplish; there might be better approaches than what you are using. – EdM Mar 31 '17 at 20:17
  • I will include a similar model tonight. Regarding your first sentence, I'm confused because I believe **here** is [Cross Validated](http://stats.stackexchange.com/) Maybe it was just a typo. After reading up on step-wise I see there is some confusion in what [step-wise] (http://stats.stackexchange.com/questions/264600/what-exactly-is-stepwise-model-selection) is. To clarify I am using AIC step-wise regression. But I will consider your advice and look into other avenues for find the "best" model/formula. @EdM – S_Stand_ring Mar 31 '17 at 23:21
  • That confusing sentence was from a mistyping. It would help if you could say more about what you want to do with your model; for example, do you wish to use it for prediction? – EdM Apr 01 '17 at 15:56
  • In the long run I do wish to use it for prediction. But from my understanding the best practiced way to do that is estimating my y-variable for times t(1-n) then doing a time series analysis. So my hope with this model is to just **estimate current pricing**. – S_Stand_ring Apr 01 '17 at 18:17

0 Answers0