Can I use Machine Learning to identify the best product configuration

Question

Is this a scenario that would work well for a ML/Pattern Recognition Model or would it be easier/faster to just filter from a large DB.

I am looking to create a system that will allow users to identify the appropriate product by specifying certain constraints and preferred features.

There are millions of possible product configurations. Lets pretend it's boxes.

Product Options:

• Size (From 1mm up to 1m) in 1mm increments
• Color: choice of 10 colors
• Material: choice of 3, wood,metal, plastic

Constraints:

• Wood is only available in centimeter units
• Red is only available in 500 mm and greater
• Wood is the preferred material
• Blue is the preferred color

So, we have 30,000 (1000*10*3) possible options. Of those, many are not viable such as 533 mm-Red-Wood

but these configurations similar to the request are possible.

• 533 mm-Red-Plastic
• 530 mm-Red-Wood
• 540 mm-Red-Wood

Notes: Our current Rules and code based tool can take anywhere from 0.5 to 2 mins to identify the preferred configuration.

We can generate a list of all possible configs and whether they are valid or not.

We estimate 30,000,000 possible configs It takes around 0.5 seconds to validate a config so with enough computing power we expect we could do 30M in a few days.

Answer 1

Can I use Machine Learning to identify the best product configuration?

Yes, if you have a set of ideal configurations from which infer what are the features that make It best.

But in your question seems that you have fixed rules and you want to generate valid configurations using just these rules. Obviously this is also possible (and easier) but It isn't machine learning since your software has nothing to "learn" from data but It has just to generate constrained configurations.