Context-Aware System Analysis And Research

In the end of the last century, personal computers (PC) becoming more popular. Calculation schematic has become more diversified. Human-computer interaction becomes one-to-one correspondence. The computers developed from company to individuals with the graphical user interface and the rapid development of multimedia technology. More and more people need to obtain information through a computer. It is a key problem we are facing that how to get context information from the massive data and use it to provide quality services for us.Firstly, this thesis introduces some knowledge about Pervasive Computing and context-aware including a frame of context-aware middleware. And it points out that ontology auto-structure and association rules mining are key step. However, traditional methods of ontology auto-structure and association rules mining don’t take context information in to consider. And this leads to the low rate of accuracy.This thesis studies the methods of ontology auto-structure based on context information, and proposes a method of ontology feature extraction called OM-OC and a similarity calculation based on context which improve the rate of accuracy. Then we propose a domain ontology auto-structure system using the new methods, and proof the superior of the system.Lastly, we make some improvements about traditional association rules mining method based on context information and users’preference and improve a new method. We can get more interesting and personalized rules through the new method. And then we proof it through an experiment.It is an important innovation point to add context information to the methods of ontology auto-structure and association rules mining. And this makes the results more accurate. This has a great significance to ontology research.