Software

Argumentation-related Essay Scoring (COLING 2016).
Demo, source code, and feature files of our experiments on argumentation-related essay scoring:

 » Online demo: Essay Scoring 

 » Java source code and feature files: EssayScoring-software-and-ARFF.zip (244 MB)

 » Readme for source code: README-EssayScoring-software.pdf (0.1 MB)

 » Readme for feature files: README-EssayScoring-arff.pdf (0.1 MB)

The essay scoring experiments are presented in a paper at COLING 2016. More information is given in the readme files.
 

Sentiment Flow (EMNLP 2015).
Source code of our experiments on the generality of sentiment flow as a model of web review argumentation. 

 » Java source code and feature files: software_SentimentFlow.zip (17.9 MB)

 » Readme for source code: software_SentimentFlow_README.pdf (0.1 MB)

 » Readme for feature files: arff_SentimentFlow_README.pdf (0.1 MB)

The sentiment flow experiments are presented in a paper at EMNLP 2015. More information is given in the readme files.
 

Shallow Argumentation Analysis (Project + COLING 2014).
Demo, webservice, source code for the usage and evaluation of our approach to domain-robust sentiment analysis, which is based on a shallow model of review argumentation. 

 » Online demo: Argumentation Analysis demo 

 » Web service: Argumentation Analysis webservice

 » Java source code: ArguAna-SourceCode.zip (32.0 MB)

 » Readme for source code: ArguAna-SourceCode-README.pdf (0.1 MB)

 » Feature files: ArguAna-FeatureFiles.zip (13.0 MB)

 » Readme for feature files: ArguAna-FeatureFiles-README.pdf (0.1 MB)

The shallow argumentation analysis is introduced in a paper to be published at COLING 2014, see » Publications. More information is given in the readme files.
 

Filtering Framework (CIKM 2013).
A software framework on top of the Apache UIMA framework that realizes our proposed view of information extraction as a filtering task, i.e., to analyze and filter only relevant portions of text in each step of an extraction step.

 » Java source code (with sample applications): IE-as-a-Filtering-Task.zip (142 MB)

 » Usage instructions: filtering-framework-readme.pdf (0.1 MB)

The filtering framework is introduced in a paper published at the CIKM 2013, see » Publications. The large file size results from a number of free code libaries included in the code. More information is given in the readme file.
 

Online Adaptation (IJCNLP 2013).
Source code for the usage and evaluation of a self-supervised online learning algorithm that adapts the schedule of information extraction pipelines to the input text at hand in order to optimize run-tim efficiency. 

 » Java source code: online-adaptation-code.zip (4.0 MB)

 » Usage instructions: online-adaptation-readme.html (0.1 MB)

The online adaptation algorithm is introduced in a paper published at the IJCNLP 2013, see » Publications. Notice that a number of free code libaries are needed to compile and run the code. More information is given in the readme file.

 

Pipeline XPS (CICLing 2013).
A protypical expert system for ad-hoc annotation pipeline construction, which can be used to automatically construct pipelines for given information needs and prioritizations of quality criteria in real-time.

 » Java source code: pipeline-xps-code.zip (5.7 MB)

 » Usage instructions: pipeline-xps-readme.pdf (0.1 MB)

The expert system is introduced in a paper published at the CICLING 2013, see » Publications. Notice that a number of free code libaries are needed to compile and run the code. More information is given in the readme file. A version of the code with all libraries can be found at the CICLING website: http://www.CICLing.org/2013/review/122-273424672657239674/cicling2013.zip 

 

Impressum | Webmaster | Letzte Änderungen am : 08.01.2017