graphinius
graphinius copied to clipboard
Generic graph (analysis) library in Typescript
- [x] remove duplicates but add a counter so we have a notion of the 'strength' of elements - [x] test input conversion - [x] add tests for set item...
Preliminary: - [ ] Define algorithms to implement - [ ] Write interfaces / classes - [ ] Write standard test cases Implement: - [ ] Disjoint Set (Or Union-Find)...
since k-step expansion on Sets seems to get REALLY slow as the periphery expands...
- [ ] research recommender loss functions & produce md / latex overview - [ ] implement & test relevant (graph-based) losses - [ ] research recommender metrics & produce...
Input: - [ ] Neo4j - [ ] Grakn.ai Output (really?): - [ ] Neo4j - [ ] Grakn.ai
Re-write Pagerank RandomWalk (maybe kick out Gauss...) to return measures that will allow iNodis to conduct performance scenarios
- [ ] target specific node types only - [x] expand on a specific subset of nodes (type e.g.) - [x] periphery @ k - [x] k-expand - [x] for...
I many algorithms, we check for all possible combinations of edges or nodes, even though only the immediate neighborhood contains interesting objects. - [ ] define a neighborhood metric -...
- [ ] construct new graph of only certain node / edge types - [ ] delete / de-activate certain node / edge types - [ ] introduce new data...