A recommendation algorithm that uses bigraphs formed by client-product interaction and meta heuristics to generate product recommendations.
This repository has 3 files here is a brief explanation of each one:
This file has the main recomendation system functionalities along with some accuracy tests, in order for the file to work you have to compile the files in the src folder which contain a cluster algorithm written in C++ , there are no extra libraries to download simply run g++:
g++ -o ARMPB src/*.cpp
This file contains the first exploratory analysis of the database which gave me informations that I used to adapt the clustering algorith in order to get a better result.
This file is a prototype of a clustering algorith based on graphs created in python, it was used as a basis for learnign the c++ algorithm.