Thread: Help Newbie With Regression

I'm very new to regression and have a ton of data I want to analyse that I think would benefit from regression but I'm not sure how to apply it.


I have 1,000 sets of rankings all 1 to 100. Each ranking has 150 variables explaining it. I want to create a model that models what is happening in all the sets of rankings.


I presume you can do a regression taking the rankings as values and the 150 variables for each ranking as explanatory variables? If so, how do I create a model that represents what is happening across all the sets?


I have completed a correlation study, by getting the correlation coefficient of each variable for each set and getting the average across the 1,000 sets.


Is there a way using regression to create an accurate model that models the rankings seen in the 1,000 sets?


Thanks in advance for your help

It depends on what sort of regression model you want. Do you want a 150-variable plane (which is basically a linear regression in 150 variables)?

Thanks yes a multi-variable regression would work if each rank was worth the same amount in each set of ranks. Unfortunately a #1 ranking in one set of ranks is not worth the same as a #1 ranking in the next set of rankings. This is due to the fact that there is varied competition for a #1 ranking across each set of ranks. Unfortunately I can think of no way to normalise the ranks into values that could be used in multi-variable regression, is there another way I can go about this?

Attach sample dataset.