Skip to content

BERT4FCA: A Method for Bipartite Link Prediction using Formal Concept Analysis and BERT

Notifications You must be signed in to change notification settings

HYYANG9/BERT4FCA-Bipartite-Link-Predcition

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

18 Commits
 
 
 
 
 
 
 
 

Repository files navigation

BERT4FCA

This is the codebase for the paper: BERT4FCA: A Method for Bipartite Link Prediction using Formal Concept Analysis and BERT

Overview

Bipartete link prediction is the task of predicting the absence or presence of unobserved links in a bipartite network. There are two types of bipartite link prediction. The first type predicts the missing links between nodes from different node sets, such as recommender system. The second tyoe predicts the missing links between nodes from the same node sets, such as predicting the newly formulated co-authorship in an author-paper bipartete network.

We propose BERT4FCA, a method for both bipartete link prediction tasks in bipartite networks using formal concept analysis(FCA) and BERT. We use FCA to extract all maximal bi-cliques and their order relations. Then we use BERT to learn the information extracted by FCA and use the trained BERT to make link prediction.

Architecture

Workflow of BERT4FCA

Files Description

Dataset

This file contains all three datasets used in the paper. BMS-POS is a product purchased transactions network provided by KDD Cup 2000.
ICFCA is an author-paper network.
Paper-keyword is an original dataset generated from the DBLP dump.

Object-Object Prediction

This file contains the codebase of bipartite link prediciton between nodes from the same node sets on each dataset.
In BMS-POS dataset, we predict two products will be likely to be bought by the same customer.
In ICFCA dataset, we predict future co-authorships or seek potential co-authors from an author-paper network at a certain time point.
In Parper-keyword dataset, we predict potentially related keywords, which may give inspiration for new research.

Object-Attribute Prediction

This file contains the codebase of bipartite link prediction between nodes form different node sets on each dataset. For each dataset, we pre-train a BERT model on objects and another model on attributes, then fine-tune on the two pre-trained BERT. We predict the potentially missing links given the known network.

Citation

Please cite our paper as:

Peng S, Yang H, Yamamoto A (2024) BERT4FCA: A method for bipartite link prediction using formal concept analysis and BERT. PLoS ONE 19(6): e0304858. https://doi.org/10.1371/journal.pone.0304858

About

BERT4FCA: A Method for Bipartite Link Prediction using Formal Concept Analysis and BERT

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published