Mult. Lin. Reg. Error Msg. -

Hello while tring to accomplish a MLR analysis for a data set consisting of X(1 - 20) vs Y1, for 165 records, I get the error message 'An error has occured with the analysis. Please make sure that your dataset meets the requirements specified in the help file.'

I've read the available program help. and cannot find the offending error. I looked at the online StatistiXL formum, and of the StatistiXL forum's 53 topics listed, at least 4 deal with the occurence of the error message, one for linear regression:

Topic Title
Topic Starter
Last Action

Cluster analysis error msgs
25th December 2005 - 17:57
Last Post by: Alan Roberts

An error has occurred with the analysis when I select more than 62 rows of data
9th May 2005 - 16:26
Last Post by: benjammin365

Error: Dataset may not meet the requirements - An error occurs with the analysis
24th March 2005 - 23:09
Last Post by: Alan Roberts

Linear Regression - Error
30th April 2004 - 08:03
Last Post by: Alan Roberts

All of your responses pretty much of either "send me the data and I'll look at it" or "what ver. of Excel and/or Windows do you use"?

I reset the format of all the numeric data in the dataset, so that the thousands separator (,) was not displayed, no change. I then tried working backwards to see when the error message first started

X20 | Y1
X19, X20 | Y1
X18, X19, X20 | Y1, etc.

The program worked fine for comparisons up to X14...X20 | Y1, and then failed with the inclusion of the X13 data.

Can you provide general things to do to avoid this error?



  • Hi Michael

    The error message you mention is the generic message that statistiXL presents when its calculations cause an error that it doesn't know how to deal with. We constantly update the list of errors that statistiXL knows about so that it can provide more informative errors but there are still a number of unknowns that fall through the net. If your dataset seems legitimate, the variables have a variance, Excels =count function indicates that it recognises all the entries as number etc then all I can suggest is that you send me your data and I will have a look at it, stepping through the code as it is analysed. This is the definitive way to track the issue and will allow us to update statistiXL to give a more informative error next time it comes across the problem.

    Best Regards

Sign In or Register to comment.