cancel
Showing results for 
Search instead for 
Did you mean: 

Eureqa -* solution notation

redenton
Linear Actuator

Eureqa -* solution notation

In Eureqa, I am getting solutions with a term "- *AE*Vsw", where "AE" and "Vsw" are variable names. What does "-*AE" mean? It doesn't seem to be the same as "- AE*Vsw" omitting "*", because the formula does not give reasonable agreement to the model in that case.

6 Replies

Hey @redenton, had a look into your question. Not sure if it is a visual bug but here for instance is a Eureqa formula like yours:

IraWatt_0-1639085113606.png

Here is the same formula displayed from the model logs:

IraWatt_1-1639085171799.png

There is now a number between the minus and asterisk. 

Model Logs can be found here under 'log':

IraWatt_2-1639085265832.png

Not sure why its not displayed in the UI.

0 Kudos

After rereading the formula it looks as though they are different. I'm not sure why the logs do not contain the formula displayed in the model

0 Kudos

It is a visual bug, here is a model with the issue:

IraWatt_1-1639089041801.png

When I export the the model:

IraWatt_2-1639089126134.png

0 Kudos

Great find! I agree that this looks like a visual bug - you may be able to see the correct expression by changing the rounding from 2 decimals to all (above the pareto chart). However, could you please send an email to support@datarobot.com to create a ticket? Provide all the information you shared here too. If you'd prefer, I can create the ticket - just let me know.

Thanks,

Jess

@IraWatt @JessLin Jess, I can create a support ticket for this. Ira's hint to look at the log was very helpful. I didn't know you could find the solutions there. Thank you Ira! Based on the log, the term in question should have been "-4.09020687809093e-7*AE*Vsw" instead of "-*AE*Vsw". Showing all digits did not fix the visual display when showing "Eureqa Solutions". Perhaps this means there is a problem in displaying a number if it is too small, like here with the e-7 notation.

0 Kudos

@redenton Great to hear the logs were the answer. Just a note that the equation in my log was different from the Eureqa equation shown (don't know if this is also a bug), I found the correct term in the Eureqa export. I also suspect it could be because its a small number : D. 

0 Kudos