Research on Patent Model Integration Based on Ontology

To integrate models and improve models collaboration, an ontology-based program-logic-combined (PLC) model representation is proposed, which is easily-implemented, reasoning-supported, has high openness and extensibility. Take the TV ontology as an example, OWL is used to describe the semantics of parameters. Based on this semantic match, model integration strategies are designed to create understandable decision problem solutions.


Introduction
Model is an important part of the decision support system.Along with the development of large decision support system, a growing need for model of integration intelligently and automatically.Nowadays,there are many methods, such as GAMS (H K Bhargava, 1997, P.193-214)proposed by Bhargava, the MIDA posted by Holocher (M Holocher, 1997, P.135-147), KVS (Kaushal Chari, 2003, P.399-413)method proposed by Chari, Byung Kwon packaging decision support module for Web service model (O B Kwon, 2003, P.375-389),but due to defficult to describe the input and output of the model, the existing model semantic properties of integration method is not ideal.How to reuse existing model program and design integration method with high expansibility, openness, automation and intelligent degree, has become an important problem to be solved for the large I 3 DSS of the integrated, interactive, intelligent decision support system.
The term "ontology" comes from the field of philosophy that is concerned with the study of being or existence.In 1993, Gruber gave a definition of ontology, namely "ontology is an explicit specification of a conceptualization" (Gruber, 1993, P. 199-220).Then Borst proposed another definition: "ontology is an explicit specification of a shared conceptualization" (Borst, 1997, P.15-19).Studer made deep research based on these two definitions, put forward the current are widely accepted definition: " ontology is a formal, explicit specification of shared conceptual models " (Studer, 1997, P.161-197).Ontology can provide structured language and explicate the relationship between different terms in order that intelligent agent can explain flexibly its meaning without ambiguity (Uschold, 1996, P.93-136).
OWL (Web Ontology Language) is a W3C (World Wide Web Consortium) recommendatory standard for ontology language, which is a language system based on description logic.OWL can be used to explicitly represent the meaning of terms in vocabularies and the relationship between those terms.OWL facilitates greater machine interpretability of web content than that supported by XML (Extensible Markup Language), RDF (Resource Description Framework), and RDF Schema (RDF-S) by providing additional vocabulary along with a formal semantics.According to the needs of different users, OWL has three increasingly expressive sublanguages: OWL Lite, OWL DL(Description Logics), and OWL Full.To capture domain knowledge from the UML class diagram as much as possible, and to ensure OWL domain ontology completeness and decidability, this paper choose OWL-DL as target language (http://www.w3.org/TR/2004/REC-owl-features-20040210).
It promotes the development of some large decision support systems, such as NED-2(Donald Nute ,2005, P.44-59), OntoWEDSS (Luigi Ceccaroni ,2003, P.785-797).But those systems are not automatic integration of model.This paper uses the idea of semantic Web services combining ontology, and puts forward the program-logic-combined model representation (PLC) method.Based on PLC representation, it gives a integrated system structure model with high expansibility, openness, automation and intelligent degree.

PLC method
Usually, there are some model representation method, such as program representation method, data representation method and logic representation method.Program representation method is easy to realize, but unfavorable to reuse model.Data representation method is favorable to realize automatic integration, but the degree of integrated intelligence is not high.Logic representation method supports reasoning, but is difficult to realize (Cai Shubin, 2009, P.713-719).In this paper the program-logic-combined model representation (PLC) method is proposed, which inherited the advantages of Program representation and logical representation, easy to implement and support reasoning.Also through the use of ontology model parameters, all kinds of model integration strategies can be designed based on the semantic matching, then create easy understanding integrated solutions.Programs using C++, Java, C computer language can be easily to package and reuse.This makes the PLC method have higher openness and expansibility.The attributes of model public interface implementation class is explained through the corresponding Model Configuration Document (MCD).MCD is the logical representation of model and written by OWL, which is made up of 3 parts--model description, model parameter and model realization.This paper choice patent as the research object, because the range of patent field is very wide, so we take the "TV" subclass of patent field for example to introduce (Chen Guohai,2007, P.20-32).Then we give an ontology model of "TV" patent a description with OWL as follows.

Presentation of model parameter
In order to further decoupling model and data, this paper introduced ontology to present the meaning of data, then used semantic to connect model with data.Experts in the field and knowledge engineer, create the Domain Data Ontology of patent field (DAO), which definite semantic concepts of the field Data.The DAO used protege to design, OWL to descript.Figure 1 is the screenshot of DAO, which set the "TV" is a subclass of "household_appliance" class.The "TV" include 17 attributes, such as applicant, appNo, abstract, appDate, patentName (Chen Guohai,2007, P.20-32).Data Administrators describe data and use DAO to create data entity, semantic concepts and their corresponding relation.
Here is a definition about a class named "TV", whose superclass is "household_appliance".The "TV" include 17 attributes, such as applicant, appNo, abstract, appDate, patentName.Then we take the "The LCD televisions with computer functions" for example to give a description with OWL referenced by the PRC state intellectual property office sites (http://www.sipo.gov.cn/sipo2008/zljs).TV ,computer parts, power and LCD part.Among them, the output of television parts, electrical connection LCD screen.TV, LCD, computer circuit board assembly in the box, saving space, reduce the cost and easy to control.</Abstract > < Name xml:lang="en"> The LCD televisions with computer functions </ Name > < Address xml:lang="en">518112 Guangdong province shenzhen longgang industrial area</ Address> < Inventor xml:lang="en"> Cai Zhuo-xian</ Inventor >

Structure of model integrated system
In this section, structure of model integrated system is briefly introduced.As figure 2 (Cai Shu-bin, 2009, P.713-719)shows, the system is divided into two modules--data and model.Data module concern data integration, data storage and retrieval.Model module concerns present management, operation and integration of model.Model integration strategy is an important part of the decision support solution.Decision support solutions include data, model and output area.Data area designated data.Model district designated the preferred integration strategy, and generating configuration document collection.The output area designated semantic concept needed by data.