Spending few weeks working on the idea/implementation of this project. Simply put the idea of this recommendation system is to implement practical models to recommend items based on users history logs. Ive used Apriori & FP-Growth which considered as Frequent pattern detection to resemble Market Basket Analysis that can address the question is the user is interested in item (X) should I recommend item (Y) or something else. This analysis is based on performnce measurements starting from the confidence AKA accuracy, Support threshold w.r.t the life metric as well. BTW Apriori & FP-Growth algorithms are consider Association Rules Mining these are heavily used by Amazon, Booking, Airlines, these platforms are built on using recommendation systems. You can check the next image to understand the different startagies that drive $$$$ to these corporates.
Additionally, I created an intro video "Recommendation Systems Tutorial (Part One)" regarding Recommendation Models, just to give familiar insights about this undrerated topic. Another advanced algorithm I've used to drive insights is Naive Bayes - Machine Learning Model which is a probability technique used to address the question "What's the probability of a user buy item (x) and item (y)?" More of measure the correlation based on the threshold setted by practitioner. Feel free to check the Naive Bayes model "Why Naive Bayes is still relevant in 2024?"
