1
$\begingroup$
  • Data about average monthly revenue from 2000 stores around whole country.
  • Gini coeff. of reve around 20%, with 50% of observation around average, very thin tails of distribution
  • Explanatory variables: geolocation of competitors, traffic drivers, purchasing power, population etc. in cachment area (most scatter plots cloud-shaped with small correlation)
  • some outlier cases

I've tried glm, svr, regression trees with boosting. All models predictions are only a little bit better then average (RMSE in avg model = 40k, RMSE in regression = 35k, reve avg = 130k), with small errors in centre and huge overestimation in left tail and underestimation in right tail.

Any idea for new model type or prediction technique that can fit better in tails without big disturbance in the center ?

$\endgroup$
  • $\begingroup$ What do you mean by over- and under-estimation? Is it that the fitted values are (greater,less) than most of the observed data, or is it something to do with variation around the predictions? $\endgroup$ – Russ Lenth Sep 11 '14 at 23:44
  • $\begingroup$ So are you saying you have 2000 datapoints ? Is average over a year or ... $\endgroup$ – seanv507 Sep 12 '14 at 6:17
  • $\begingroup$ @rvl : My current models predict small revenue cases to have much bigger reve then actual (in opposite to high revenues). $\endgroup$ – user147787 Sep 12 '14 at 7:21
  • $\begingroup$ @seanv507 : I have 2000 different stores and for each average monthly revenue from last year. $\endgroup$ – user147787 Sep 12 '14 at 7:23
  • $\begingroup$ Do you mean 2000 x 12 months or 2000 $\endgroup$ – seanv507 Sep 12 '14 at 7:26
-1
$\begingroup$

try Random Forest using R

As a random forest is a meta estimator that fits a number of decision tree classifiers on various sub-samples of the dataset and use averaging to improve the predictive accuracy and control over-fitting Random Forest!

| cite | improve this answer | |
$\endgroup$
  • 5
    $\begingroup$ This answer is more of a comment, please expand. $\endgroup$ – Xi'an Apr 8 '15 at 11:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.