Authors:
Haytham Mohtasseb
;
Amr Ahmed
;
Amjad AlTadmri
and
David Cobham
Affiliation:
University of Lincoln, United Kingdom
Keyword(s):
Commonsense knowledge base, Semantic network, Ontology development, Psycholinguistic, Text classification.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Data Engineering
;
Enterprise Information Systems
;
Information Systems Analysis and Specification
;
Knowledge Engineering and Ontology Development
;
Knowledge Representation
;
Knowledge-Based Systems
;
Networked Ontologies
;
Ontologies and the Semantic Web
;
Ontology Engineering
;
Ontology Matching and Alignment
;
Ontology Sharing and Reuse
;
Symbolic Systems
Abstract:
PsychoNet 1 has demonstrated the feasibility of integrating psycholinguistic taxonomy, represented in LIWC, and its semantic textual representation in the form of commonsense ontology, represented in ConceptNet. However, various limitations exist in PsychoNet 1, including the lack of concluding context of the concept annotation. In this paper, we address most of those limitations and introduce a new enhanced and enriched version, PsychoNet 2. PsychoNet 2 utilizes WordNet, in addition to LIWC and ConceptNet, to produce an integrated contextualized psycholinguistic ontology. The first and the main contribution is that, in PsychoNet 2, each concept is annotated by the potential (most representative) contextual psycholinguistic categories, rather than all applicable categories. The second contribution is the enrichment of LIWC through utilizingWordNet. This in fact produced an enriched version of LIWC that may also be used independently in other applications. This has contributed to sub
stantial enrichment of PsychoNet 2 as it facilitated including additional number of concepts that were not included in PsychoNet 1 due to lack of corresponding words in the original LIWC. A sample application of text classification, for a mood prediction task, is presented to demonstrate the introduced
enhancements. The results confirm the improved performance of the new PsychoNet 2 against PsychoNet 1.
(More)