Search

Aiming at the status that various electronic text materials are increasing rapidly, This work brings forward a model of automatic classification of electronic text information in order to manage and use these text information effectively: the algorithm of segmentation of word based on word dictionary and statistics, preprocessing of text, design of weight function of feature words and collecting them, expression of text vector space, latent semantic indexing and clustering algorithm of text, etc. The experiment has proved that the model had satisfactory classification effect as well as high calculation and storage efficiency.

With the wide application of the distributed workflow management system, the load balancing policy that can improve system's performance is being emphasized by people. A load balancing model, load balancing machinery, is presented in this paper, and it is introduced into the distributed workflow management system. And the system structure of the model is dealt with in detail. The load balancing policy is also analyzed. At last the model is realized in CORBA.

The association rule mining is an important research subject of knowledge discovery. Aiming at the common method of mining for attributes of quantitative type in database, we analyze the existing defects and put forward a method of applying fuzzy set theory to association rules mining. Due to the problem that each attribute's importance is different in specific purpose mining, we put forward a solution by assigning corresponding weight to attribute of different importance. Based on this idea, we put forward a mining algorithm using fuzzy weighted association rules and through the given experiment we testify the feasibility of the algorithm, and point out the existing defect of the algorithm demanding improvement in future.