Modeling Concept-Attribute Structure
نویسنده
چکیده
We apply hierarchical Latent Dirichlet Allocation (hLDA) to the problem of ontology annotation; automatically extending WORDNET with new concepts and annotating existing concepts with generic property fields, or attributes. The resulting annotations are evaluated along two dimensions: (1) the precision of the ranked lists of attributes at each concept, and (2) the specificity of the attribute assignments to WORDNET concepts. We find that hLDA and several variants outperform previously proposed heuristic methods, significantly improving the resulting annotation quality, and confirming the ability of topic models to reduce extraction noise.
منابع مشابه
Linguistic Characteristics of Nouns Representing the Concept of “Power” in American Political Discourse
The article analyzes the linguistic nature of some nuclear lexemes (power, control, authority, influence) verbalizing the basic concept of the English-language political discourse – the concept of “Power”. The paper reveals the structure of the concept, its conceptual features, as well as the main characteristics of the concept representatives at the language and functional levels.
متن کاملOn Concept Algebra: A Denotational Mathematical Structure for Knowledge and Software Modeling
Concepts are the most fundamental unit of cognition that carries certain meanings in expression, thinking, reasoning, and system modeling. In denotational mathematics, a concept is formally modeled as an abstract and dynamic mathematical structure that encapsulates attributes, objects, and relations. The most important property of an abstract concept is its adaptive capability to autonomously i...
متن کاملConcept design of Vehicle Structure for the purpose of computing torsional and bending stiffness
Automotive design engineers face the challenging problem of developing products in highly competitive markets. In this regard, using conceptual models in the first step of automotive development seems so necessary. In this paper, to make a body in white conceptual model, an engineering approach is developed for the replacement of beam-like structures, joints, and panels in a vehicle model. The ...
متن کاملComparative Study between Traditional and Heuristic Approach on Close-degree Concept Lattice for better Attribute Reduction
Concept lattice is a new mathematical tool for data analysis and knowledge processing. There is an increasing interest on application of concept lattices in the different information systems. The concept lattice may be used for representation of the concept generalization structure generated from the underlying data set. Attribute reduction is very important in the theory of concept lattice bec...
متن کاملFrom Preferences over Objects to Preferences over Concepts
We present an approach to preference modeling and learning preferences from data based on formal concept analysis. We consider techniques to derive preferences over attribute subsets from preferences over objects, including ceteris paribus preferences.
متن کامل