regression fit
• I have a set of data that contains 6 inputs and 1 output (about 500 rows). Basically, I'm trying to figure out the complex relationship between the 6 inputs (say different component of concrete mix, i.e. sand, aggregate, water etc.) in order to predict the strength of concrete (output).The data was collected through a number of experiments. I'm trying to figure out the relationship between them. For instance, if you reduce water volume, the strength decreases.

I was wondering if anyone can provide me with some tips

 Tweet
• Hi mzn,

It sounds like the best way to start would be to set up your target expression to search for the output (strength of concrete) as a function of the other 6 columns. ie, your target expression would look something like:
concrete_strength = f(sand, aggregate, water, ...)

The model you get out of that will distill the most important variables and relationships that affect concrete strength. You can find more details on setting up your target expression here: http://formulize.nutonian.com/documentation/eureqa/user-guide/set-target/

Best,
Jess
• Thanks Jess.

One more thing, is there a way to ensure I get an equation with the simplest form possible.
• Also, if I know that the equation is of a polynomial form (sand^2+aggregate^0.2 etc.), how can I force Eureqa to look for a function with this feature?
• Hi mzn,

Eureqa focuses on finding the simplest form of an equation possible, and returns a front of the best models across a set of complexities. You can then go through the models that Eureqa has found and focus on the models with the best accuracy/simplicity tradeoff for you.

You can also use the target expression to fit coefficients to a known form - see the "fitting coefficients" portion of the target expression documentation: http://formulize.nutonian.com/documentation/eureqa/user-guide/set-target/#target-expression.

Best,
Jess
• Thanks again Jess. One last thing, if I have this situation,

1) Temp. is a function of Time, b, d and V, so that Temp. = F(Time,b,d,V).
2) Then, k is a function of Temp., Basically, k equals a unique value at each Temp.
3) Finally, Deflection is a function of Time, C, G, and S*k

Is there a way to incorporate these three items into one target expression? Such that, I can find out what is the deflection at any time through one equation.
• Hi mzn,

You could try a target expression with multiple expressions like this:
Deflection = f0(Time, C, G, S*f1(Time,b,d,V))

You can then go into the Reports tab and select the "Print current results details" report to see the separate functions for f0 and f1.
• Thanks, but does that account for the relationship between k and Temp.? The relationship between k and Temp. has a polynomial form. I know that equation too. The idea is to get Temp. based on Time, b, d and V. Then, this Temp. will be input to the polynomial form to get k. Then, k*S, C, G and Time will be used to arrive at the deflection.

#### Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!