Web Service Discovery Based On Conceptual Ontology Similarity Computation

In this more and more informational society, web is related to people's daily life and give big convenient to people's life. In the future, all the software will be put out as service.After purchase the right to use the service, people could use all kinds of service through Internet Explorer. Shortly to say, web service is a series of function program to achieve particular aims. So, how to discover the service on internet have been a key problem for public to concern with. The appearance of web service is a way to resolve this problem.In the past years, what we did in service match maker is based on key words, this kind of match maker has high precision. But it is difficult for this kind of match maker to find all of this right service, and it is hard to finish the operation between web services. In this year, sematic web have been put forward, it is give out a chance to resolve this kind of problem. Combining semantic web with web service as sematic web service, it is a good manner to fix now's web service problem.During the process of sematic web service discovery, how to calculate the similarity is a key step. The result is good or not have a serious inflection of the result of service discovery, even the service combination in the next step. This paper focus on semantic web service matchmaker algorithm,calculate the similarity between service and the system framewrok.In this paper, using ontology to calculate the similarity between two services. Discuss the calculating algorithm and the position where all kinds of service property will on, proposes a composithive similarity algorithm and a out-first service discovery frame:1. Proposing a composithive similarity algorithm, combining more detail matchmaker algorithm.In order to calculating similarity between two concepts, we think about not only semantic distance,depth, density, but also the share information. First, check the relationship between two concepts, if there has disjoinwith relationship between two concept, we will do nothing later,the similarity calculate is stop. Otherwise, check the subsume relationship between two concept, divide the matchmaker level,then, calculate similarity according to the level.calculate the share information similarity between two concept, if the similarity is 0, stop. Otherwise, calculating the similarity based on distance. At last, compose the two similarity as the final similarity.2. Give out the result of experiment result of the discovery with composithive similarith algorithm,in this experiment,input,output all used as filter service,and give the contrast result of this algorithm between classic algorithm.through the result we can saw this algorithm have a between result at recall/precision.3. Divide the service discovery process into filter and sort two parts. And have a different define of input during the match making process, use input to sort services. To prevent the low recall for the reason of the influence of user can not exactly know what the input will be to find out particular output.4. Use SME2 to test the matchmaker algorithm and the similary calculate algorithm. And give the result campare with the classic algorithm of precision, recall, memory use, and query response time. Analyse the reason of bad results, and give out the way to resolve that kind of problem.