This repository contains the LVA code for our paper "Linked Variational AutoEncoders for Inferring Substitutable and Supplementary Items", published in Web Search and Data Mining (WSDM) 2019
But I have noticed that you have done some preprocessing steps on the raw dataset. Could you please upload a sample dataset for us to run? like your dataset_graph_4class_filtered.json file? Or could you please give us a brief description of how to filter out the unrelated items?
BTW, what is the vector dimension when you do the Doc2Vec?