Extracting Meronymy Relationships from Domain-Specific, Textual Corporate Databases

Abstract

Abstract.Various techniques for learning meronymy relationships from opendomain corpora exist. However, extracting meronymy relationships from domain-specific, textual corporate databases has been overlooked, despite numerous application opportunities particularly in domains like product development and/or customer service. These domains also pose new scientific challenges, such as the absence of elaborate knowledge resources, compromising the performance of supervised meronymy-learning algorithms. Furthermore, the domain-specific terminology of corporate texts makes it difficult to select appropriate seeds for minimally-supervised meronymylearning algorithms. To address these issues, we develop and present a principled approach to extract accurate meronymy relationships from textual databases of product development and/or customer service organizations by leveraging on reliable meronymy lexico-syntactic patterns harvested from an open-domain corpus. Evaluations on real-life corporate databases indicate that our technique extracts precise meronymy relationships that provide valuable operational insights on causes of product failures and customer dissatisfaction. Our results also reveal that the types of some of the domain-specific meronymy relationships, extracted from the corporate data, cannot be conclusively and unambiguously classified under well-known taxonomies of relationships.

Conceptual Representations for Computational Concept Creation

“…Broadly speaking, semantic relations are labeled relations between meanings, as well as between meanings and representations. In contrast to associations, semantic relations have meanings as Hyponymy-hypernymy Hearst 1992;Caraballo 1999;Kozareva et al 2008;Pantel and Ravichandran 2004;Pantel and Pennacchiotti 2006;Snow et al 2005;Navigli and Velardi 2010;Bordea et al 2015 Meronymy-holonymy Berland and Charniak 1999;Girju et al 2006;Ittoo et al 2010;Pantel and Pennacchiotti 2006 Causal relations Khoo et al 2000;Girju and Moldovan 2002;Blanco et al 2008;Ittoo and Bouma 2011 indicated by their labels. The number of semantic relations is virtually unlimited.…”

Towards an Automatized Generation of Rule-Based Systems for Architecting Eco-Industrial Parks

“…Ranked list of synonyms [21] Similar to finding synonyms, a number of algorithms have been proposed for identifying part-whole relationships. [25] presents an approach for finding part-whole relationships in domain-specific data. Part-whole relationships in domain-specific data is more difficult to address than such relationships in general language, as the available data is much smaller.…”

“…As a consequence, it is more difficult to train machine learning algorithms. The approach proposed in [25] leverages on part-whole relationships extracted from an open-domain corpus and extends these relationships by domainspecific relationships. [26] is arguably the first automatic part extraction from a large, unlabeled, corpus with a reported precision of 0.55.…”

AUDD: Audio Urdu Digits Dataset for Automatic Audio Urdu Digit Recognition

“…Currently, some work studied ASR for Urdu [13][14][15][16][17][18][19][20]. Additionally, numerous works [12,17,[21][22][23] have been done on spoken digit recognition for Urdu.…”