Skincare Recommendation

What it does:

Our recommendation engine will provide The Ordinary product recommendations based on a skincare profile (the user can input their concerns and optionally their skin type and skin sensitivity levels to better optimize the product recommendation) ensuring that all product conflicts (contraindications) and potential adverse reactions are avoided. The engine is built so that the most efficient/most targeted product possible is recommended over any other products - we have limited one type of product (as of now, we have water-based and moisturizer) to one specific concern for ease of use.

Common Sense Aspects

Ensuring no conflicts, ensuring only the least number and most efficient (through our common sense data in database.pl) products are given. It can also go off of minimal inputs (just concern). It will also ensure that for highly reactive products like Glycolic acid, it ensures that all skin type/sensitivity level matches whereas for unreactive products like Alpha Arbutin, less checks are run.

Files:

database.pl: Justifies the logic behind the 'rules' and defaults/constraints/order of precedence based on knowledge regarding the efficiency of products. Our program 'ranks' certain products in order of efficiency and will only recommend a lower efficiency product in the case that all products of higher efficiency are unable to be recommended because of constraints or if they pose any contraindications with other recommendations.
recommendation.pl A series of rules which allow for conditions for recommendation(), don'trecommend() and contraindications(). These rules were built using the database.pl constraints/hierarchy of efficiency as a reference, which used The Ordinary's product information.
sampleinput.txt How the sample input file will be provided by the user.

Note every products output will have at least one water based product and one moisturizer (rose hip oil or vitamin c, depending on constraints).

sample_products.txt: How the program outputs the recommendations.
sample_products_and_models.txt: How the program outputs the recommendations with models.
minimal_input.txt
minimal_output.txt Showing how the program can extrapolate information.

What we learned:

How to work with Prolog and s(CASP). Building recommendation engines. Planning out the logic behind rules.

What's next for Skincare Recommendation System:

Add more products for the anhydrous/emulsion portion to generate a complete layering guide. Ensure a better output which generates a 'schedule' for the recommended layering guide/regimen. Special thanks to the mentors who were able to help us make this possible!