Hyper Optimizasyon Model or Which model ?

First of all, thank you very much First of all, thank you very much. I have a project in which I am trying to examine the effect of fertility parameters on profitability in dairy cattle farms.

I have about 10 different criteria and these criteria are age at first calving, calving interval, lactation duration and so on. then we know the milk yield, calf yield and what the profit is due to the effect of these parameters.

What I am trying to do is this point. I want to find the effect of changes in the input parameters on the profit. So if the lactation period decreases by one day, what is the profit. I want to build a model where I can do this for the other 10 variables.
When I looked into it, I saw that I can use the hyper optimization model.
What other methods can I use. Which example can guide me.

Thanks in advance

Are you trying to optimize milk production, calf yield or both?

Yes of course i would like milk optimiza

but Also in case change the inputs , i would like to see how and how many change the output same time

Hi if you create a multiple linear regression model the coefficients would give you an indication of the effect one input parameter has
By Hyper Optimization you mean Hyperparameter optimization? If so that would be used in order to “tune” a model to get the best kind of predictions. That is not what you search for (if I understand correclty)
br

@hakanimamoglu could you upload a dataset that has your task? If this would not spill any secrets.

I think you will need some sort of Optimization with some restraints. But that is only a first guess.

1 Like

anket için denemeler.xlsx (47.3 KB)

it is not normal values because i did not make survey and did not take data from program. i filled random but care about interval. yellow colones inputs ande 2 red colones outputs. For example M2 Cell value = 17 , if M2 cell =16 Colon , How can T2 ,U2 cell change. and other inputs
Thank You

Could you possibly provide the column headers in English?

anket için denemeler.xlsx (47.7 KB)

Of course

I’ve taken a preliminary look at your data and have a number of questions. It seems to me that you can’t optimize total yield directly since it depends on the number of cows. You should focus on average daily yield. Having said that the data doesn’t make much sense, e.g. look at the following rows:
No Cows Ave. Yield Total Yield
1 173 38 117800
7 5 39 230958

I have a number of questions which can be found in this Excel workbook:
Questions.xlsx (9.3 KB)
Finally, I’ve looked at the correlation between Ave. Yield and other factors and its very low.

You could use H2O AutoML, it will do the hyperparameter optimization for you, builds multiple models and does validation using crossvalidation. Here’s an example:

2 Likes

I don’t think he needs (or wants) hyperparameter optimization. He’s after variable optimization/correlation (I think.) Its hard to tell and he hasn’t responded to my questions. Reread his original post.

you right. no corelasyon with your number because i create this datas randomize.
But i want to learn how can i analyz with correct datas.

Take a look at these workflows. They should help you determine the most important regression coefficients. As @mlauber71 said if you want real optimization you’ll probably have to employ some Python code. As @evert.homan_scilifelab.se commented you can try AutoML which can help find the best model. To use it effectively you’ll need to play with setting hyperparameter bounds.

1 Like

@hakanimamoglu if you want an optimization with constraints it will be good to have a real dataset if you are interested in problems of dairy farming indeed. This article seems to deal with such problems but does not provide data or code:

Application of a mathematical framework for the optimization of precision-fed dairy cattle diets
https://www.sciencedirect.com/science/article/pii/S175173112300318X

This is the original study referred here:

If you want to train optimization in general you might have to look for good sample datasets where you could use Optimization and root finding (scipy.optimize) — SciPy v1.11.3 Manual (for example).

You can also try to employ ChatGPT though in the past I had problems to get it to write consistent code and samples for such problems. Maybe there are useful plugins now.

2 Likes

Thank you for reply.
i m trying and i m working to learn.

AutoML Regression and Classification Examples

At above examples i should replace examples excel files with my excel files when i prepare excel

Thank you very much for advice

@hakanimamoglu if you want more regression machine learning examples I can offer this collection but we mentioned depending on your task it might not give you the answer you are looking for.