Completeness And Knowledge Acquisition On Incomplete Information System

Nowadays, with the rapid development of Database technology and network technology, Data and Knowledge in all fields increase dramatically. And Data and Knowledge become more uncertainty with the involvement of human, as well as the relations between them become more complex. How to mining the hidden, useful, valuable information from so much uncertain, messy, under- disturbance Data and Knowledge, is a very serious challenge for the mining of useful information, so development and exploration of methods for information processing is in urgent. The paper tracking the international research status, mainly research the completeness of IIS and the related knowledge acquisition models, and the results obtained not only enrich improve incomplete information system handling, but also are expected enormous applied value due to the widespread applied background of these methods.The main results and originalities are outlined below:1. Introducing fuzzy relation into incomplete information system and building an information completing platform, we presented two different completeness methods for incomplete information system. The platform provide effective data supporting for the completeness of IIS based-on multi-partitions and that of IIS with various partitions. Compared with some traditional methods, the methods may not only a new try for completeness of IIS, but also may obtain a more reasonable and effective results.2. To avoid the necessity of reasoning on data with missing attribute values, we propose a method of data decomposition. The original incomplete data is decomposed into data subset without missing values and that with only missing values. Next, the concept of completeness template is defined. Then the method of completeness is accomplished on the complete information subsystem generated from completeness template.3. By introducing Formal Concept Analysis Theory into incomplete formal context (a special form of IIS), we mainly discuss the tolerant concept of incomplete formal context and the acquisition method of tolerant rules based on it. In order to compress large-scale set of tolerant rules, we develop an inference rule to reduce the redundant tolerant rules, and finally obtain a complete and non-redundant set of tolerant rules. Users can take this method into various practical needs through setting parameters. Example proves the method is effective.