{"id":28966274,"url":"https://github.com/dice-group/lpbenchgen","last_synced_at":"2025-06-24T07:33:00.339Z","repository":{"id":45157452,"uuid":"348849838","full_name":"dice-group/LPBenchGen","owner":"dice-group","description":"OWL explainable structural learning problem Benchmark Generator","archived":false,"fork":false,"pushed_at":"2022-09-19T12:58:51.000Z","size":2069,"stargazers_count":1,"open_issues_count":4,"forks_count":0,"subscribers_count":8,"default_branch":"main","last_synced_at":"2024-07-30T20:49:47.756Z","etag":null,"topics":["abox","benchmark","concept-learning","negative-concepts","ontology","owl","positive-concepts","tbox"],"latest_commit_sha":null,"homepage":"https://dice-group.github.io/LPBenchGen","language":"Java","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"agpl-3.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/dice-group.png","metadata":{"files":{"readme":"README.md","changelog":"CHANGELOG.md","contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2021-03-17T20:56:31.000Z","updated_at":"2021-12-15T10:42:51.000Z","dependencies_parsed_at":"2023-01-18T14:16:50.933Z","dependency_job_id":null,"html_url":"https://github.com/dice-group/LPBenchGen","commit_stats":null,"previous_names":[],"tags_count":3,"template":false,"template_full_name":null,"purl":"pkg:github/dice-group/LPBenchGen","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dice-group%2FLPBenchGen","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dice-group%2FLPBenchGen/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dice-group%2FLPBenchGen/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dice-group%2FLPBenchGen/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/dice-group","download_url":"https://codeload.github.com/dice-group/LPBenchGen/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/dice-group%2FLPBenchGen/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":261627810,"owners_count":23186694,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["abox","benchmark","concept-learning","negative-concepts","ontology","owl","positive-concepts","tbox"],"created_at":"2025-06-24T07:31:15.301Z","updated_at":"2025-06-24T07:33:00.309Z","avatar_url":"https://github.com/dice-group.png","language":"Java","funding_links":[],"categories":[],"sub_categories":[],"readme":"# About\n\nLPBenchGen is an OWL explainable structural learning problem Benchmark Generator.\n\nExplainable AI is becoming more important and more and more systems to tackle that problem are developed.\nHowever, there is a lack of benchmark datasets.\n\nLPBenchGen tries to tackle that problem by providing a tool that can generate benchmarks using a TBox and an ABox.\n\nThe ABox can either be behind a SPARQL endpoint or in an RDF file.\nUsing an RDF file allows an Open World Assumption. This is restricted in SPARQL, which generally allows only a Closed World Assumption.\n\nHave a look at our latest Documentation at [https://dice-group.github.io/LPBenchGen/](https://dice-group.github.io/LPBenchGen/)\n\n\n## What is a Learning Problem in this context?\n\nA learning problem is a set of positive examples and negative examples, whereas one example is an Individual contained in the ABox.\n\nA system then may be benchmarked to find the best concept/class expression or try to find all other positive Individuals .\n\n## How does it work?\n\nLPBenchGen will create a lot of theoretically class expression based upon the TBox and checks if these have enough solutions using the ABox.\nThe user can set parameters allowing longer and more complex as well as smaller and easier expressions, how many of these should be created and a lot more.\n\nUsing these expressions LPBenchGen creates positive examples for one learning problem.\nHowever, it is not straight forward to retrieve negative examples.\n\nFor example if we have a big diverse Database and have a concept such as `Tree and (hasLeaves some Leaves)`:\nA negative Example such as `Elefant` wouldn't be useful at all.\nWhat would be useful is something like `WindFlower-1` as it has leaves, but is not a Tree.\n\nHence, LPBenchGen creates a few class expression based upon the positive class expressions in that matter.\nThese basically negate some parts of the original expression and hence providing useful negative examples.\n\n\nAfterwards these concepts will be used to retrieve positive and negative examples and thus creating the learning problem.\n\nA benchmark in this context is a set of such learning problems.\n\n\nFor a detailed description please have a look at Underlying-Model\n\n\n## What does LPBenchGen generate?\n\nLPBenchGen generates a training benchmark set, a test benchmark set and  a gold standard for the test benchmark.\n\n\n## My ABox is huge, no current system can work with that!\n\nNo worries. LPBenchGen can create a small ABox fitting the problems in the benchmark.\n\n## How do I evaluate my system?\n\nIf you generated your benchmark accordingly (see Usage)\nLPBenchGen provides a small evaluation. Simply use the test benchmark, the gold standard, and the actual system answers and evaluate. See Usage for the correct format and eval.\n\n## How to download\n\nDownload the latest release jar file at [https://github.com/dice-group/LPBenchGen/releases/latest](https://github.com/dice-group/LPBenchGen/releases/latest)\n\nSee Getting-Started for how to execute the generation.\n\n## Where can I find the code?\n\nThe code is open source at [https://github.com/dice-group/LPBenchGen](https://github.com/dice-group/LPBenchGen) and you can code with us if you want to :)\n\n## Where do I submit a bug or enhancement?\n\nPlease use the Github Issue Tracker at [https://github.com/dice-group/LPBenchGen/issues](https://github.com/dice-group/LPBenchGen/issues)\n\n\n## Documentation\n\nHave a look at our latest Documentation at [https://dice-group.github.io/LPBenchGen/](https://dice-group.github.io/LPBenchGen/)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdice-group%2Flpbenchgen","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fdice-group%2Flpbenchgen","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdice-group%2Flpbenchgen/lists"}