Construction Of Steel Ontology And Design Of Instance Population System

There are rich data resources in materials science. In particular, amounts of materials data resources resided in open materials web sites are available. However, most of these data resources are shown in the form of HTML tables, and it is difficult to distinguish the attributes and values or find semantic information in tables for machine because of the semi-structured feature of HTML tables. Therefore, in order to integrate these heterogeneous data resources reside in tables in semantic level, it is helpful and necessary to create a domain ontology and achieve the population from these data resided in tables to target ontology, which is also one of the main tasks of materials informatics.The thesis proposes a steel semantic model(named STSM) based on ontology and logic rules for the representation of the steel knowledge in level of semantics. STSM is developed with the consideration of the features of materials data and the developed process is presented. We describe the content and organization of STSM which covers the basic knowledge in steel domain. Domain axioms and logic rules are also designed to enhance the reasoning ability of STSM. Then, we choose metal materials data in HTML tables as research object, and the methods for materials data extraction from HTML tables and the instance population to steel semantic model STSM are designed through analyzing available materials web sites. In table data extraction, based on sibling comparison, a method for materials knowledge extraction from HTML tables is proposed. We find the sibling tables from sibling documents through similarity between two tables, and then use FRFC(i.e., the First Row matching and First Column matching) strategy to identify table pattern of sibling tables based on sibling table comparison. Based on the mapping from tables to table pattern, a table object can be divided into one or more simple tables whose attributes are located in first row or first column so that these data in simple table can be extracted. The extracted data is mapped to the predefined schema which will facilitate the population to materials ontology. Moreover, based on the mapping from predefined schema to target ontology STSM, the population from metal materials instance data to STSM is achieved. Further, we give the experimental evaluations and prototype, and the steel knowledge can be integrated in semantic level by our approaches.