Development of an Ontology-based Short-Term Loan Expert System

• Main Authors: Cheng-yi Chang, 張成儀
• Other Authors: Li-yen Shue
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/33618884752627811280

碩士 === 國立高雄第一科技大學 === 資訊管理所 === 94 === Abstract This research investigates the feasibility of developing an ontology-based expert system for determining short-term loan application. Evaluation of loan applications requires knowledge from multiple sources, and is basically a semi-structured job. The main information sources to support the evaluation processes normally consist of financial statement, financial ratios, and other non-financial information. Because of its complexity, it takes experienced accountants to do the job well. Researchers have applied expert system technology to assist accountants. In general, the expert’s decision knowledge consists of domain knowledge and operation knowledge. The former represent the basic knowledge of financial statement, which may not vary due to different person and conditions. The later represents the application logics of the decision processes, which may vary depending on different level of experiences. Traditional way of designing knowledge bases usually lumps all knowledge into a rule base. Its major defects are: (1)difficulty in modifying knowledge bases, (2)difficulty in sharing knowledge with other systems. In view of this, this study applies Part-Whole Relationships to analyze knowledge content and build their knowledge structure. We apply ontology technology to represent domain knowledge, and uses rule-based technology to represent operation knowledge. We use Protégé as the knowledge platform to integrate the two modules. The interface part of the system is developed with Java language that is compatible with Protégé and JESS. The loan domain knowledge of this study could be applied to different systems and cases, and achieve the purpose of knowledge reuse and sharing.