Research on Construction of Agricultural Domain Knowledge Service Platform Based on Ontology

. Scientific researchers’ increasing demand for knowledge service under the new situation, makes it urgent to embed information service into user research process, ad build an incorporate knowledge platform that integrates knowledge, skills, tools, and services of certain professional field. This paper put forward the technical solution of agricultural domain knowledge service platform based on ontology, including resource organization based on ontology, platform design and development. The construction progress of ontology base and service functions based on ontology are shown by application practice in rice domain.


Introduction
Scientific and technological innovation is the source of power that promotes social development and progress. In the era of knowledge economy, innovation and application of knowledge has become the most important part of scientific and technological innovation. Any research on scientific and technological innovation is inseparable from the support of science and technology information resources, and the construction of e-science environment.
Scientific users have an increasing needs [1,2] for service integration, knowledge management, knowledge services, communication and collaboration under the new situation. This makes it become urgent to build a research and knowledge service platform to integrate knowledge, technology, tools, services of certain domain, so as to embed information services into users' research process, achieve service innovation, and support the scientific and technological innovation.
Research on the construction of domain knowledge service platform has made significant progress internationally. Lots of countries are paying more and more attention to domain knowledge service platform, and some large, national level research projects and practical activities have been carried up. There are some systems or software tools that we can refer to, such as VIVO based on ontology [3,4], VRE based on SOA [5,6], Harvard Catalyst, Sciologer of Columbia University [7], SKE of CAS [8], the open source software of virtual learning environment SaKai, [9][10][11], etc. These construction experiences provide a good foundation for this study.
Ontology is a good technical tool which shares common understanding of the structure of information among people or software agents and enables reuse of domain knowledge. This paper put forward the technical solution of agricultural domain knowledge service platform based on ontology, including resource organization based on ontology and platform design and development. Current construction and application situation in rice domain was introduced at the end.

Resource organization based on ontology
Resources organization is the root of a knowledge service system. Knowledge organization system construction is foundation and core of realizing the transformation from scientific literature information services to knowledge service. Agriculture itself is a complex subject, so what needs to be solved in this system is the organization of core knowledge content. Using ontology which can represent the meaning and relevance of knowledge more accurately, laid the foundation for construction of the knowledge service application.
Regarding domain knowledge service application as the guidance, advanced international mainstream knowledge organization technology should be adopted in agricultural domain knowledge organization. Existing knowledge organization materials were used for reference in ontology design and ontology instance base construction.
Ontology of agricultural domain knowledge service platform consists of scientific research ontology and domain ontology.

Scientific research ontology
Scientific research ontology reveals and reflects research elements and their mutual connection, which is the foundation of knowledge discovery. VIVO, put forward by Connell University, is very useful reference. VIVO integrated with powerful ontology management tools, through construction of ontology around scientific experts ("people"), using Jena inference system to realize associated navigation and retrieval of research objects [12,13]. Considering scientific research agent, research condition, research activities and research output are core concepts during scientific research progress, in this paper, we proposed the scientific research ontology around research progress. Its main class resources are as Fig.1. Domain ontology can focus on the whole domain, or a small part of it, or even be a combination of several disciplines. According to the general ontology framework, extending ontology instances facing specific area is the key of content construction of agricultural domain knowledge service system. Therefore the domain ontology of this platform will be constructed after the target application domain is identified.

Platform design and development
Agricultural domain knowledge service platform is a personalized information platform for professional researchers, and also an integrated information system which is Is based on

Responsible for
Belong to Journals the organic integration of library services, users' research process and the internet service environment.

System architecture
This paper presents the 4-layer architecture of agricultural domain knowledge service platform, as shown in Fig.3, which includes support layer, resource layer, application layer and user layer.   The underlying support layer aims at the building of hardware and software environment which is the technical foundation of the entire system.  Data service interface mainly achieves the scalability and extensibility of agricultural domain knowledge service platform, supporting data exchange between the platform and third-party systems or services.

Functional structure
 The system basic functions such as system management, user and privilege management, and statistics support the system's normal operation.

Platform development
According to the characteristics of large volume data of the system, distributed architecture was used, so as to benefit sharing system computation and storage pressure.
Java was the main development language, based on J2EE architecture, and MyEclipse was the development tool, with the environment of "Windows2008+ Tomcat6.0+Mysql5.5".
During construction of the knowledge base, O-R (ontology relation) mapping technique was applied to achieve dynamic storage of resources, and Hessian technology was used to obtain the cross language, cross platform and more openness of the knowledge base.
By adopting the SSH structure, knowledge application of the foreground system could dynamically adjust columns and information project from the knowledge structure, to support cross domain application. Knowledge service functions were realized by us- And knowledge organization and management functions in background system were implemented by "generic desktop" technology, which provides Windows operating experience under the Web system.

Application practice in rice domain
Application to a specific domain is the "last mile" to realize construction target of agricultural domain knowledge service platform. Combined with country-level major projects, according to the discipline construction and development status of Chinese academy of Agricultural Sciences, we selected rice as the target domain and deployed the rice domain knowledge service system, which provides knowledge service to scientific workers engaged in rice research. Fig.5 gives the rice knowledge service system homepage screenshots.

Service functions based on ontology
In rice domain knowledge service system, besides the ordinary functions including domain knowledge retrieval, knowledge navigation, knowledge acquisition, and maintenance of knowledge organization system, there are two typical service functions based on ontology.
First is visualization of research entities and their relationships. As shown in Fig.6, scientific researchers, research topics, institutions, literature, and relationship between them are revealed in one "knowledge map". then the platform provides basic information of the disease or insect pest, as well as related symptoms, occurrence position, rice growth period, methods of prevention, drugs, and relevant literature; another is that, with some symptoms input by users, the platform presents suspected plant diseases and insect pests after reasoning based on ontology, so as to give treatment suggestion and relevant literature links (Fig.7).

Conclusion
Domain knowledge service platform is not a new service model, but its construction is still in the exploratory. The practice we have tried using ontology in agriculture domain knowledge service is very meaningful. However, there are still a lot of issues that need further improvement and solvent in the future, such as core services, construction of domain ontology, data standards, users' research process and their work environment, non-technical factors, interdisciplinary knowledge exchange and so on.
In short, building an integrated knowledge service platform for professional areas is a general trend, but carrying out the work in a large scale and showing the effectiveness still need a period of time. That needs constant exploration to new technologies, new methods of knowledge services to continuously promote technology innovation.