{"id":20549545,"url":"https://github.com/akullpp/elac","last_synced_at":"2026-06-06T15:31:57.794Z","repository":{"id":15302557,"uuid":"18032336","full_name":"akullpp/ELAC","owner":"akullpp","description":"Ensemble Learning for Anaphora- and Coreference-Resolution-Systems","archived":false,"fork":false,"pushed_at":"2014-03-27T15:11:19.000Z","size":21252,"stargazers_count":3,"open_issues_count":0,"forks_count":1,"subscribers_count":3,"default_branch":"master","last_synced_at":"2025-01-16T16:42:49.479Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Java","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/akullpp.png","metadata":{"files":{"readme":"README","changelog":null,"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":"2014-03-23T12:00:26.000Z","updated_at":"2023-09-16T21:23:14.000Z","dependencies_parsed_at":"2022-07-31T04:38:01.001Z","dependency_job_id":null,"html_url":"https://github.com/akullpp/ELAC","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/akullpp%2FELAC","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/akullpp%2FELAC/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/akullpp%2FELAC/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/akullpp%2FELAC/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/akullpp","download_url":"https://codeload.github.com/akullpp/ELAC/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":242153814,"owners_count":20080566,"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":[],"created_at":"2024-11-16T02:18:24.386Z","updated_at":"2026-06-06T15:31:57.754Z","avatar_url":"https://github.com/akullpp.png","language":"Java","funding_links":[],"categories":[],"sub_categories":[],"readme":"License: Apache\n\nUsage Instructions\n\n\nPlease read INSTALL for installation instructions.\n\nImportant notes:\n\n1. BART is a very complex tool with very specific options. Therefore, you have to\nset it up and run it on the corpora specified in ./config/config.xml as\nPathToTrainingFiles and PathToTestFiles with your individual options.\nFurthermore, we suggest that you train BART itself otherwise its results may be\nto bad to be considered by the classifiers.\nThis is a process our program does not want to do automatically since many\nexternal factors and the user's intention must be considered when running BART.\nBy running this program, we assume that you did the necessary steps with BART\nand the output of BART is located in the markables folder.\n\n2. We are sorry for the huge console output. This is a general problem of MMAX2\nand was reported by us. A concise output of the results will be printed after\nthe specific process has finished.\n\n3. Please read about our findings in the ./poster.png\n\n\nInstructions:\n\nStart the CLI via \"java -Xmx1024M -jar ELAC.jar\"\n\n(1) starts the JavaRap preprocessing unit.\nThis is a necessary prerequisite for the training (4) and testing (5) process!\nIt creates a javarap subfolder in the Basedata folder specified by\nPathToTrainingFiles/PathToTestFiles in the config.xml with the JavaRap results.\nPlease select whether you have JavaRap installed locally (1-2) or you need to\naccess it via an SSH connection (1-1). You will be prompted for your\ncredentials if you choose to run JavaRap via SSH. It is very important that you\nforward X in order to get prompted for your credentials.\n\n(2) starts the feature distribution analysis tool.\nThis is a standalone tool. Please specify whether you want the feature\ndistribution of the test corpus (2-1) or training corpus (2-2) which have to be\nspecified by PathToTrainingFiles/PathToTestFiles in the config.xml.\nAfter the process is finished, you will find a file in the ResultOutputDir (see\nconfig.xml) with the total/average distribution of each feature in the selected\ncorpus.\n\n(3) starts the Salsa converter.\nThis is a standalone tool. It converts an XML file in Tiger/XML format which\nyou have to provide to an XML file with an inline annotation that can be\nprocessed by MMAX2. It will be written to the folder specified by\nResultOutputDir in the config.\n\n(4) starts the training of the classifier.\nYou will need the results from the JavaRapPreProcessing unit (1) in order to run\nthis process!\nFurthermore, you will need to set following options in your .config/config.xml:\n\t\n\ta) PathToTrainingFiles: Please enter the path to the training corpus.\n\t\n\tb) Runner: Please enter the ACR-Systems you want to use. In order to use an\n\tACR-System it must be specified in de.uniheidelberg.cl.swp.testacr.\n\tCurrently BART, JavaRap and LingPipe are supported. Please feel free to add\n\tyour ACR-System.\n\n\tc) EvaluationMethod: MUC6 is the default to choose. If you consider to use\n\tDIRECTNEIGHBORSONLY please read the extended documentation.\n\t\n\td) FeatureFilter: Enter the features which are to be extracted. The features\n\thave to be specified in de.uniheidelberg.cl.swp.featureExtraction.features.\n\tPlease note that every feature which is specified in FeatureFilter has to\n\thave its specific entry with its custom value or value range.\n\nThe results of each individual ACR-System and the generated results.arff file\ncan be found at the location specified by ResultOutputDir in the config. The\nresults.arff is used in the testing process (5).\n\nNotes:\n1. Multiple options like Runner or FeatureFilter as well as the specifc entries\nfor each feature have to be separated by \";\".\n2. Depending on the size of the corpus, this process might take a long time.\n\n\n(5) starts the testing/machine-learning process.\nYou will need a WEKA compatible ARFF file in order to run this process, which is \neither created by the training process (4) or can be provided in form of a \ncustom ARFF file (5-1)!\nFurthermore, you will need to set the following options in your .config/config.xml:\n\n\ta) PathToTestFiles: Please enter the path to the test corpus.\n\n\tb) classifier/subclassifier: Specify the classifier/subclassifier. We\n\tcurrently support:\n\tJ48 as classifier and subclassifier\n\tBAYES as classifier and subclassifier\n\tBAGGING as classifier which needs a subclassifier\n\tADABOOST as classifier which needs a subclassifier\n\tNEARESTNEIGHBOR as subclassifier for stacking\n\tZEROR as subclassifier for stacking\n\tKSTAR as subclassifier for stacking\n\tBFTREE as subclassifier for stacking\n\t\n\tc) stacking: True or false. If you want to use stacking, you will have to\n\tspecify a classifier for level 1 and a subclassifier for level 0.\n\t\n\td) options: Options for the (sub-)classifier. Please read the WEKA\n\tdocumentation for an extensive list.\n\nThe result of the testing process can be found at the location specified by\nResultOutputDir in the config.\n\nNote, depending on the size of the corpus, this process might take a long time.\n\n\n(6) starts the Ablation Testing tool.\nThis is a standalone tool. You will have to specify the name for the new file,\noutput will be written to the folder specified by ResultOutputDir in the \nconfig. Furthermore, you will have to specify the correct path to the test \ncorpus in the PathToTestFiles entry in the config.\n\nNote, depending on the size of the corpus, this process might take a long time.","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fakullpp%2Felac","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fakullpp%2Felac","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fakullpp%2Felac/lists"}