期刊名称:International Journal of Computer Science & Technology
印刷版ISSN:2229-4333
电子版ISSN:0976-8491
出版年度:2013
卷号:4
期号:2
页码:727-730
语种:English
出版社:Ayushmaan Technologies
摘要:Annotation of web pages is an area of research which is getting lot of attention as the count of websites of specific topics and as a whole is increasing very fast. Since all the databases are accessible over web through HTML representations and data extraction over web is becoming more and more dynamic. Such data is huge and for applications such as online shopping comparison, article collection etc. Annotation of such collected information leads to several advantages including fast decision making, relevant information visiting, to reduce the time of futile searches, historical data management and elimination of older searches. This paper is intended to provide an insight of the annotation techniques and application of few techniques to provide the required results with the above stated advantages. Works of various researchers in the field of annotating data has been more on limited tokens and focus is on creating dynamic annotations only. This work proposes to apply dynamic annotations on web sites data with tokenization done using all sort of tokens including long text having no specific tokens. For machine learning and training frequency based annotations, common knowledge annotators and schema value annotators are being applied which are going to facilitate for correct annotation process. For annotation website pages shall be looked for content type, presentation style, data type, tag path and adjacencies of the contents.
关键词:Data Annotation;Web Databases;Data Alignment;Data Filtering; Frequency Annotation;Multimode Text