Research On Biomedical Ontology Matching Algorism Base On Field Ontology

Ontology, it is a concept origins from philosophy, in the file of computer science and information science, it can help a lot in many problems in knowledge egineering, so it had been well developed in recent years. The biomedical field is also have a rapid speed been developed in recent years, and the ontologies, due to its structural, are applied widely in this field. And more and more organizations, groups, researchers are using ontology to convinient their work, and build a lot of ontologies, but there is not a very classify stadard of how to build an ontology, so different ways can result many different ontologies, ontologies base on one thing can be very different due to the way how people build it, and we call these different ontologies heterogeneous ontologies, in order to connnect the application base on those ontologies, we should find a way to match and alignment those h eterogeneous ontologies, it has a great meaning in the computer science field. And s o on, to alignt two biomedical ontologies base on two different biosource can also have a great meaning on disease predict.To alignt two ontologies, the most important step is to match them, this step is to caculate the similarity between the concepts or r elations in two ontologies by many method, then combine these similarity by their weights to get results, and then filter these results to find the final mappings between two concepts or relations.In this paper, we mainly focus on the specific of the biomedical ontologies, proposed that using biomedical field ontologies' synonymous and cross reference relation informantion to find more similarity informations between concepts. Meanwhile, according to that the names of the concepts mostly are specific, so we caculate the frequency of different words, then weight these different words base on their frequency, in that case, we will find less false matching result which could be found in other mathing method. Then, because the biomedical field have a lot of disease information, we can build connections between ontology and these information, to find more relations between concepts. Finally, we modify the weight based on different method, let the string and field resourse value more because it's more important, and then select the results.After we get the maping information, the next step is to alignt the ontologies, if we need to use different ontologies, we can use the maping information to connect, but it's not convienient, so to merge two ontologies into one ontologies can deal it, so the last part of this paper discuss how to build a merged ontology based on two ontologies and the maping information of these two ontologies.Finally, this paper base on the research have done, developed a tool to merge ontologies, we can use two heterogeneous ontologies as input, or two ontologies base on two things but on one type as input, to get output that the mapping informations of the two ontologies inputed and the merged ontology base on those two ontologies.