A domain ontology construction method for intelligent application

Abstract

The invention discloses a domain ontology construction method oriented to intelligent application. Firstly, multi-domain data related to intelligent application is obtained. Identify and refine the requirements of intelligent application, and construct ontology-oriented intelligent application requirements analysis model; Define the concepts and their attributes in ontology, the relationships of each type, create instances, and encode them formally; Then Jena is used to implement domain ontology-based reasoning to complete the missing information. Secondly, the architecture of ontology is evaluated by using multi-dimensional quantitative index, and whether it is returned or not is judged to modify ontology. Finally, the support degree of ontology to intelligent application requirements isevaluated, and the architecture is updated dynamically to respond to the changes of application requirements until the ontology can meet the application requirements. The invention adopts the circulating development structure, emphasizes the interaction between the intelligent application requirements and the cross-domain ontology construction in the whole life cycle of the ontology, is suitable for representing the cross-domain knowledge association oriented to the intelligent application, and solves the problem that the existing ontology construction method is insufficient to support the intelligent application.

Description

Technical field [0001] The invention belongs to the technical field of computer applications, and specifically relates to a field ontology construction method oriented to intelligent applications. Background technique [0002] With the revival of the artificial intelligence boom, the process of intelligence in various fields has accelerated. The realization basis of intelligent applications is a comprehensive and accurate representation of domain knowledge, so that the machine can process and understand domain knowledge. Therefore, how to enhance the ability of knowledge representation to support intelligent applications is an important research topic in the field of intelligence. Ontology is a clear formal specification of the shared conceptual model. It is a knowledge sharing tool with strong representation ability, machine understanding and support for reasoning. It can provide a good conceptual foundation for the realization of intelligent applications. Therefore, it is used...

AI Technical Summary

Problems solved by technology

[0004] The analysis shows that the existing ontology construction methods ignore the driving effect of smart applications on ontology construction and the support ability of ontology to smart applications, and cannot guarantee that the constructed ontology can be implemented and meet the final application requirements.

First of all, the requirements analysis in the construction process is too simplified, and neither the determination of the scope nor the selection of resources can fully express the rich connotation of intelligent applications, let alone quantify the application requirements for machine understanding and implementation

Correspondingly, in the evaluation process, it is difficult to accurately measure the fit between the ontology and the requirements of intelligent applications, to optimize the application functions of the ontology based on the evaluation results, and to dynamically update the architecture to respond to changes in application requirements.

In addition, the existing methods all focus on the selection of concepts and the establishment of their architecture, and most of them use the hierarchy of concepts as the main line to organize domain knowledge. The number of non-categorical relations is small and their role in the ontology structure is weak. It can better reflect the static knowledge system but it is difficult to describe complex knowledge associations, and the realization of intelligent applications needs to discover laws and grasp the situation from the associated knowledge, which is difficult to achieve only relying on hierarchical relationships

[0005] Specific to the field of anti-fraud, the traditional method is based on simple filtering and interception based on blacklist or business rules, which is difficult to deal with complex and changeable fraud methods

Images