A Role of Snippets in Document Search in the Web - Assignment Example

As time goes by and new technologies emerge, more and more ontologies are constantly developed for the semantic web. Ontology is constantly developed that is deemed useful to companies. Development of ontology can be based on the design approach from a high set of reusable ontology. While designing the ontology is necessary, the primary concern is how the search results reflect the true nature of the search query (Stoilos et al 2005). The aspects that determine the snippets generated are based on the query input, therefore for this to be realized proper design and development of relevant ontology need to be done. Snippets are important in the determination of the content held by the resulting document. In order to avoid the burden of opening each document, which can list millions of search results, short information, is provided with regard to document metadata or location. Labels can be associated with every document while on the other hand provision of the main topic of the document or metadata can be done. Snippet generation systems help generate the information that users find important in the course of their searches.

Snippet generation is a special type of extractive document summarization in which some information fragments of the total information are chosen and displayed based on the search query match. It can also be referred to as query-based summarization. Initially, the search engine presented the first part of the document as a snippet in the results. The information to display was based on some number of kilobytes of the total document (Sakai, & Spark-Jones, 2001). These snippets were not however representative of the total document, nor were they related to the search query in most cases.

Few algorithms appear in the literature concerned with the snippet generation effectiveness.  However, Google has tried in the past and was the first to produce query-based summaries snippet generation. In the past, most experimental work has focused on the comparison of the values of searches in relation to the other types of summarization. Silber and McKoy (2000) describe a linear time lexical chaining algorithm for use in generic summaries but they offer no empirical data for the algorithm's performance. White et al (2006) in their research on snippet generation demonstrate some advanced research on the WebDocSum system.