Semantic Search Engine for Daily Work Reports Integrating Heterogeneous Ontologies

In recent years, Knowledge Transfer is one of the problems of business management. Companies are required to create a mechanism to transfer knowledge of experts to the next generation. Submitting work reports is one of examples of efforts to convert tacit knowledge to explicit knowledge. There are many companies letting their employees submit work reports after completion of their work. However, most of data is just accumulated, but it is not fully reused. Factors obstructing utilization of work report data are below; First, data of work reports is enormous and there are scattered descriptions concerning important business knowledge so that it is impossible to acquire business knowledge by reading work reports efficiently. Second, since work reports are plaintext, their structure cannot be seen so it is impossible to access the necessary information immediately. Therefore, in this research, we constructed a system that supports knowledge transfer by utilizing work report data. Specifically, we constructed domain ontologies that define the meaning of terms related to business knowledge and performed information extraction on work reports by using the domain ontology and sentence pattern matching. Then, based on the result of information extraction, a set of texts including business knowledge was selected. Finally, we constructed a system that provides business knowledge to users by using work report data as a knowledge source.