1.1.1. Research Background
The current era is the era of network information technology; the speed of internet development has already exceeded people’s initially expectation. The amount of network information has increased dramatically with the widespread popularity of the Internet. Work and study on the Internet has become part of people's life, amazing amount of data of learning resources in the form of electronic documents widely spread on the Internet. Such a huge amount of information makes the Internet to be a good source of information. But how to provide more convenient for the network users a simple and efficient way to retrieve, it is often more difficult.
The currently popular two search engines - Google and Baidu are two preferred tools for users to retrieve information from the Internet. According to the survey report, as of June 2015, China's Internet users has more than 650 million, the Internet penetration rate has reached 48.8% in the mass of information access, the search engine accounted for more than 80%(The 36th Statistical Report on Internet Development in China. 2015).
The search engine provides a convenient way for Internet users to search for useful information, but the traditional information retrieval system is often based on keyword and full-text search technology, It can only be based on the user's keyword input matching search, rather than attach importance to information semantic mining, The consequences of doing so is to return to a large-scale information, but also the lack of a certain relevance, it also cannot give users a satisfactory result. Although the current search engine is supplemented with various techniques for improving accuracy, the following problems still exist due to defects in the model (Chen, 2009; Li, 2009; Gao,2014; Yang, 2006):
The first: the problem of “information overload”. Excessive pursuit of high recall leads to returning a large number of search results. It is unlikely for the user to give a keyword that exactly matches the content in the source. So, users often need to spend more time on searching for useful information;
Second: the problem of “faithful expression”. Due to the different requirements, although different users use the same keywords to search, but their respective emphases are not necessarily the same. Currently widespread search engine ignores the user's browsing behavior and their personal interest characteristics. Therefore, the user's needs cannot fully performance in the search process, and the search results are often unsatisfactory;
Finally: the problem of “isolated island of vocabulary”. At present, keyword-based retrieval is only a literal match, but vocabularies in the human brain does not exist in isolation, How to extend the query content through semantic understanding and association based on the user's query vocabulary becomes a difficult problem.
Traditional search engines exposed above problems become major factors impede the retrieval performance boost. Researchers in this field are attempting to exploit the new perspective to change this unfavorable situation, such as: We can use the semantics to re-weave, store and retrieve the Internet information from the perspective of semantics. The application of ontology reasoning and other related technologies can make the search technology to get rid of the semantic representation limit, and then deep into the semantic level, and then improve the level of retrieval from the user's point of view.
1.1.2. Research Significance
When semantic retrieval techniques are applied to analyzing the retrieved objects and requests, we can no longer confined to the keyword and the full text of the surface meaning, but also take the key words into a deeper level with the concept of semantics from the perspective of semantic understanding.
The concept and architecture of the Semantic Web was first introduced in late 2000 by Tim Berners-Lee. Ontology in the field of semantic retrieval technology has an irreplaceable position, Ontology is the development of cutting-edge knowledge representation technology in the information technology field, It not only can analyze the intricate relationship between concepts, and then get its abstract nature, still can achieve logical reasoning through a certain degree of development tools, and use the relevant knowledge of graph theory to form the corresponding concept hierarchy. Because Ontology's formal information representation is more complete, It allows search engines to concepts within the domain, axiomatic knowledge and the relationship between different concept to form a unified understanding. All of the above advantages make the whole massive information in the search process can achieve the semantic level of support, and make information resources positioning more accurate and more efficient, and greatly reducing the probability of redundant information, so it can improves the recall and precision of the retrieval system. Research on Ontology-based semantic retrieval technology can improve the accuracy and accuracy of current network information retrieval, which is of great significance for academic theory, scientific research and practical application.