A metaheuristic approach wrapper generation for web data alignment and annotation from web databases

Abstract

Today World Wide Web (WWW) is the major all kind of newlineinformation store and has been very successful in providing information to newlineindividuals. The web has become the ideal medium for several database newlineapplications and these applications stores data in large databases that newlineuser s query, access and update via the web. The web databases return newlineseveral search results dynamically on the web browser, these search results newlinecontains deep web pages in the Hyper Text Markup Language (HTML) page newlineformat. Every web page comprises multiple Search Result Records (SRR) newlinerelated to the query. These search results have to be extracted and newlinemeaningful labels have to be assigned. In this research work, automatic newlineannotation approach has been presented that aligns the data units on the newlineresult page into various groups, each group contains data having the same newlinesemantic. Each group is then annotated from various aspects and these newlineannotations are aggregated to predict the final annotation. Then, annotation newlinewrapper is automatically constructed for the search site and is utilized to newlineannotate new results from the same web databases. In this research an newlineunsupervised learning methodology is applied to generate wrappers for the newlineweb page. In this research work, an improved wrapper generation using newlinemetaheuristic techniques has been proposed. The term metaheuristic newlinegenerally refers to approximate algorithms for optimization that are not newlinespecifically expressed for a particular problem. Some Metaheuristics are newlineinspired by natural processes such as evolution, others are extensions of newlineless sophisticated algorithms such as greedy heuristics and local search. In newlineapproximate methods such as Metaheuristics it sacrifice the guarantee of newlinefinding optimal solutions for the sake of getting good solutions in a newlinesignificantly reduced amount of time. Metaheuristics are usually easier to newlineimplement than classical gradient-based techniques.