DiverSE Coffee: Marcel Heinz

Abstract

We are happy to receive Marcel Heinz as a guest speaker for our next DiverSECoffee. Marcel is a research assistant in the software languages team at theUniversity of Koblenz-Landau. He is working towards a Ph.D. His generalinterests are in Software Language Engineering, Modeling and Ontologies. Hisresearch focus is on assisting at the comprehension of software technologies.Enters Marcel: Wikipedia is a rich source for classifying entities in manydomains. Wikipedia's category graph focuses on classification combined withsecondary uses of categorization. Wikipedia's infoboxes define entity propertiesthat also relate to classification. In this paper, we aim at validating and,ultimately, improving the classification data available on Wikipedia. To thisend, we develop a multi-dimensional methodology which combines availablefragments of ground truth, diverse feature extraction, and metrics computationin a systematic manner. We apply the methodology to a case study which resultsin a strongly validated classification of software (computer) languages. ThePresentation is held this Friday in Markov room at 10 am.

Date
Event
DiverSE coffee
Location
Rennes, France

We are happy to receive Marcel Heinz as a guest speaker for our next DiverSE Coffee.

Marcel is a research assistant in the software languages team at the University of Koblenz-Landau. He is working towards a Ph.D. His general interests are in Software Language Engineering, Modeling and Ontologies. His research focus is on assisting at the comprehension of software technologies.

Enters Marcel: Wikipedia is a rich source for classifying entities in many domains. Wikipedia’s category graph focuses on classification combined with secondary uses of categorization. Wikipedia’s infoboxes define entity properties that also relate to classification. In this paper, we aim at validating and, ultimately, improving the classification data available on Wikipedia. To this end, we develop a multi-dimensional methodology which combines available fragments of ground truth, diverse feature extraction, and metrics computation in a systematic manner. We apply the methodology to a case study which results in a strongly validated classification of software (computer) languages.

The Presentation is held this Friday in Markov room at 10 am.