No abstract

The rise of eXtensible Markup Language (XML) as a standard for data exchange and representation has made a lot of interest on querying XML documents that conform to an ordered tree-structured data model. It involves labeling, which assigns specific strings to XML documents. Existing schemes are based on natural order, lexicographical order or Variable Length Endless Insertable Code (VLEI) order, which needs relabeling. A novel order concept called vector order has been introduced, which is the basis of the dynamic labeling schemes. Vector order is a simple, yet most effective solution to process updates and it avoids relabeling during updation takes place. Existing dynamic XML labeling schemes, suffer from the complexity introduced by their insertion techniques even if there is less/no update. To make improvent the application of vector order has been proposed to both prefix-based labeling scheme and range based labeling scheme. However, all the existing techniques have high update cost, they cannot completely avoid relabeling in XML updation, it will increase the size of the label which will influence the query performance. By applying the vector order the label generation time, label size and querying time may be reduced during updation takes place. Hence it will improve the performance of the query system.

No abstract

Designing well structured websites has long been a challenge to facilitate effective user navigation. The reason for poor website design is that the web developers understanding of website structure is considerably different from that of the users. The reorganization of a website structure changes the location of familiar items. They would not be frequently performed to improve the navigability. The existing Mathematical Programming model facilitates user navigation on a website with minimal changes to its current structure. It is appropriate for informational websites whose content are static and relatively stable over time. It allows a page to have more links than the out-degree threshold if the cost is reasonable. Number of relevant mini sessions decreases as path threshold increases. Thus the Mathematical Programming model focuses on enhancing the design of existing links before adding new links.To improve the navigability further, an enhanced model using Graph Partitioned Clustering algorithm is proposed. It is used to group the potential users with similar navigation pattern from preprocessed web log data. Clustering results include number of visits made to a single webpage, most frequently viewed page and navigation behavior of the users. To minimize the average weighted shortest distance between pages and the average surfing distance, the enhanced mathematical programming model is proposed and applied to the clustered results. After identifying the existing links to be improved and new links to be added, the web site current structure has to be reorganized. Thus, it will improve the user navigation on a website effectively.

No abstract