Hybrid approaches for the analysis of relevant high quality xml web data

Abstract

In recent days, the eXtensible Markup Language (XML) based web newlineapplications are widely used in data exchange and network services. In machine newlinelearning database, keyword search can be implemented and also it is possible on newlinegraph structure that combines the relational, html and XML data. The search newlinemechanism for XML files is a very important and essential technique to retrieve newlinethe text content from XML. But, it is difficult to identify user intentions through newlinethe keyword. In web search engine, keyword search is one of the most important newlinesearch representations for regular users. For this reason, XML language is newlinebecoming a standard in web data representation. XML supports keyword search newlineand it also allows users to create queries without the knowledge of query newlinelanguage and database schema, so that it is considered as a user-friendly method. newlineThe user can access the relevant web data by analyzing keyword in XML web. newlineIn web search engine, querying and extracting data from web has been an ongoing research issue since the birth of the web. As amount of data is increasing day by day, extracting data becomes a difficult task. In order to solve this issue we develop an efficient method for relevant XML web data quality newlineanalysis. In the initial research, the quality analysis of relevant XML web data is newlinedone using clustering and classification technique. Clustering is employed by newlineModified Fuzzy C Means (MFCM) clustering and classification by K- Nearest newlineNeighbor (KNN) algorithm. At first, a number of XML documents are collected newlineand clustered based on keyword depending on type of XML files by means of newlinemodified fuzzy c means algorithm. In order to find the relevant XML web data, newlinethe clustered features are then applied to the KNN classifier which results in newlinehigh accuracy. newline newline