Wrapper Extraction and Integration using GNN

Abstract

Extracting data from the web is most prominent and discussing field now days. Extraction of useful semi structured data from the World Wide Web is the main aim. The extraction from the large web normally known as deep web is done by form submission cannot be done by any ordinary search engine. In data mining the automatic detection and extraction of data becomes bulky due to the uncertain structures of websites. Data extraction techniques developed till date are normally dealing with the extraction of text, audio, video etc. but there is a little and bit weak methods regarding the extraction of image data is the concern of recent research. One of the arts of image data extraction is DOM Document Object Model, it is a solution to extract the semi structured data but by the time the HTML documents are getting larger and contain more data. It is found that there is getting lengthy processing time and also emerged with noisy information. In the given research work we have tried to give a graphical representation of for the improvement of Wrapper Extraction of Image using DOM and JSON (WEIDJ). We have proposed the Graph Neural Network (GNN) to be used in wrapper extraction to improve the performance.