I am looking for a model for a suggestion engine. Problem: a new user comes in to system with 20 images, I need to find out all subsets that has 2 to 5 images and then score each set. Score will be between 1 (user will 100% like it ) to -1( user will 100% not like it). My data set contains thousands of images sets. A different earlier users declared liked (1) or not like(-1) for the image sets in my data set. I need a model such that when I present 2-5 images it will give me a score between 1 and -1.
Bu iş için 34 freelancer ortalamada $546 teklif veriyor
I’m working on the same project newspapers. I recommend n news to each user. I’m using instruments are similar with matrix factorization and gradient boosting to design recommendation system.