How to implement a data model for generating a dimensional grid?

Question

Machine learning

Task: to implement a model for determining the size range for tailoring.
Problem: for the correct preparation of patterns for sewing, a change of 2-4 cm in 6 parameters at the same time is critical, those 6 parameters-sizes form a dimensional grid for sewing (Neck circumference, chest coverage, back length, dog height at the withers, waist circumference, distance between front paws), there is no single classification of dog sizes and they jump from breed to breed, the condition of the dog and age also affect. In order to create a 3D model and properly prepare patterns, you need to define groups of sizes, those parameter links for one set of patterns and form a dimensional grid for transfer to 3DClo and Graphis
There is: a form for collecting data on the site that takes into account 12 parameters of the dog, there is a typology of breeds (breed groups with similar proportions), while more than 500 size values ​​\u200b\u200bare accumulated by breed groups, but you can quickly collect this data from dog owners on the site
Need to: Understand how implement a model that will be able to:
1. Determine from the entered data what size of the grid the dog belongs to
2. Supplement the access values ​​for the size, taking into account the difference in values ​​(typify the size grid)
3. Determine which of the existing models (parameters for the final clothes are set) are suitable for this dog

To understand in your own language: in the dimensional grid for a person, models are typified taking into account race and national parameters, there are dimensional grids for Europeans, Asians, and so on, which under 3 parameters Height, chest coverage, waist coverage form patterns, since the length of the sleeve or the length trouser leg is calculated from the already accumulated data and these formulas are used in 2d graphics and 3d graphics models. For dogs, such information is not collected, the data is scattered, and you need to create a model that will first collect this data, segment it by a mass of values, and, based on the result, will be able to form universal sizes that will be used in patterns

Question: how to build such a model, what can be used and on what to implement

Answer 1

The question is very interesting, but sorry, not for the forum. Just not for this format: it’s impossible to answer three paragraphs, it’s pointless to write a scientific treatise here.
In addition, there is no ready-made solution, here you need to experiment. The general direction is certainly clear - machine learning, classification and clustering. But the final choice of method (or - most likely - a sequence of methods) and "model" - requires a fairly deep analysis.
I don't really understand your last sentence. Ready-made solutions are unknown to me personally. Well, if you set a goal, then maybe something can be found somewhere, but most likely - for human clothes, so you still have to adapt and not weakly. Therefore, you can use any machine learning libraries. Most traditionally - from the Python ecosystem. There are a lot of things that you will have to tailor the final decision from.
In general, the project is interesting. There will be specific questions - ask, we will be happy to help.

Answer 2

In my opinion, everything is simple. You need to find a specialist who made such dimensional grids for people. Let him do the same for the dogs.
The business decided that a programmer, and even with "machine learning" would do it better? Or will it cost less? In my opinion, this is a serious mistake.