Splunk Search

How to calculate a linear regression for each field and predict the next possible number in my search query?

saenito
New Member

Suppose I have this data (but in thousands, is just an example):

ID class qty
1 cup 5
2 cup 6
3 cup 2
4 cup 7
5 mug 1
6 mug 3
7 mug 4

I want to calculate a linear regression for "qty" (or just to use the predict command) for each "class", so I want one predict (or linear regression can work too) next possible number for "cup" and for "mug".

Considering that there are not only 2 different classes but thousands of it, how can I do this?

ps: I have tried with "map" command but it limits my results to only 10 iterations even after I have modified the "maxsearches" parameter but didn't work.

0 Karma

skoelpin
SplunkTrust
SplunkTrust

I agree with @martin_muller .. This is a BIG job for the predict command and will kill your hardware and will most likely not give you the desired results. Download the MLTK and fit a model with linear regression and it will give you feedback as to how accurate it is

0 Karma

saenito
New Member

Thanks, i'm trying with MLTK, altough i can not see how to make one "prediction" for each type

i can only get the models to show data for "total" numbers

0 Karma

martin_mueller
SplunkTrust
SplunkTrust

I'm not quite sure I understand what you want, but it sounds like a job for the Machine Learning Toolkit, not for the predict or even map commands: https://splunkbase.splunk.com/app/2890/

saenito
New Member

yes, the problem i keep finding in splunk is that regardless i use Machine Learning Toolkit it only works for a single series, and i want to do the calculation for each class, thats why i have tried using map to "iterate" trough ecach "class" or series and the making the predict command

and i see is the same with machine learning toolkit, it only works for only one thing, for example: how many sales im expecting for next time slot, but not how many sales im expecting for each item

thanks anyway

0 Karma

niketn
Legend

@saenito, in that case, instead of Forecast Time Series you should check out Predict Categorical Fields

____________________________________________
| makeresults | eval message= "Happy Splunking!!!"
0 Karma

niketn
Legend

@saenito, you should definitely explore the Machine Learning Toolkit App which comes with several Machine Learning Algorithms not just Linear Regression.

If you just want to implement Linear Regression, try to check out the following Splunk Wiki: https://wiki.splunk.com/Community:Plotting_a_linear_trendline

____________________________________________
| makeresults | eval message= "Happy Splunking!!!"
Get Updates on the Splunk Community!

Index This | I am a number, but when you add ‘G’ to me, I go away. What number am I?

March 2024 Edition Hayyy Splunk Education Enthusiasts and the Eternally Curious!  We’re back with another ...

What’s New in Splunk App for PCI Compliance 5.3.1?

The Splunk App for PCI Compliance allows customers to extend the power of their existing Splunk solution with ...

Extending Observability Content to Splunk Cloud

Register to join us !   In this Extending Observability Content to Splunk Cloud Tech Talk, you'll see how to ...