recommenders-team/recommenders

Which unsupervised method can infer explicit interaction from occurance(count) data?

Open

#1,384 opened on Apr 24, 2021

View on GitHub
 (0 comments) (0 reactions) (0 assignees)Python (2,972 forks)batch import
help wanted

Repository metrics

Stars
 (17,706 stars)
PR merge metrics
 (Avg merge 6d 16h) (10 merged PRs in 30d)

Description

Our input data is a table of edges (edge weight is implicit interaction, which weighted with co-occurrence, edge type was also implicit which inferred from prior knowledge) and a list of nodes (two types: users and items). Group-truth is a small set of edges that are known to exist. The ground-truth set is too small to train a model, therefore, we want to do unsupervised analysis. For example:

edgetable.csv (potential edges) dst | src | attr A | B | edgetype1 A | B | edgetype2 A | C | edgetype1 B | D | edgetype1

nodetable nodeID | type | sample1(count) | sample2(count) | sample3(count) A | nodetype2 | 5 | 6 | 8 B | nodetype1 | 3 | 5 | 9 C | nodetype1 | 2 | 4 | 6 D | nodetype2 | 5 | 2 | 5

ground-truth B D edgetype1

Our purpose is twofold:

Whether there is an explicit interaction between two types of nodes turely exist in given samples?
How to determine the edge type?

Thank you for your help！

Contributor guide

Research direction
Explore unsupervised graph inference methods such as network embedding (e.g., node2vec, GraphSAGE) or matrix factorization (e.g., SVD) to predict explicit interactions from co occurrence data. Consider using co occurrence counts as features for link prediction. For edge type determination, consider multi relational graph embedding methods like ComplEx or DistMult.
Tech stack
None
Domain
machine learningdata
Issue type
Research
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-2 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Fresh
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
PythonMachine Learning
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
50