Ontology-based Agricultural Vertical Search Engine For Arid Areas

With the rapid development of the Internet, the agricultural information has alsoincreased sharply. However, the efficiency of agricultural users obtains relative information islow in the massive complex data ocean. General search engine can not meet the requirementsof agricultural users in information retrieval precision, the current exists’ agricultural searchengine have a series of problem such as the lacking of credibility agricultural service platform,the low rate of agricultural search engine put into practice and so on. In order to solve theseproblems has mentioned above, this paper has do some research on agricultural informationvertical search engine model, which is specialized for arid region of northwest China on thefoundation of regard the western arid agricultural IOT(Internet of Things) informationplatform as credibility agricultural service platform. The main contents and conclusions are asfollows:(1) Research an agricultural domain ontology building method based on all kinds ofagricultural literatures. In order to improve the automatic degree of establishing ontology andreduce the dependence on experts, web data extraction technology is used to obtain the fieldof corpus, and the corpus in the field is segmented and cleaned for automatic extracting ofdomain concepts, correlation analysis and hierarchical clustering is used for discovering therelationship between the domain concepts. Lastly, results are fixed artificially to establish theontology. Experiments demonstrated that this method improves the automatic degree ofestablishing ontology and guarantee the performance of all methods.(2) Research an agricultural information vertical search engine based on ontology. Inorder to improve the precision of arid agricultural information retrieval, design andimplement an agricultural information vertical search engine based on ontology. Firstly,realized the collection and filtration of network information on the foundation of ontology,implemented the functions such as link filter, information obtains, page analysis and topicrelevance judgments in orders by combine topic crawler with ontology to collect and filterarid agricultural information, and then indexed the relevant pages with the strategy of theinverted index. Secondly, the ontology-based information retrieval is completed. Combinedquery expansion technique with agricultural domain ontology to expand user’s search queryon the host, parallel and the next three modes, ordered the results through the level of the relevance, and then recommended user the relative words according to the relationshipbetween the next word and the host word. Experiments showed that, the agriculturalinformation vertical search engine based on ontology can improve and optimized theaccuracy of search result.(3) Designed and implemented the agricultural information vertical search engineplatform based on ontology. In order to make it easily to manage and use the agriculturalvertical search engine and ontology established above, this paper designs a platform based onB/S+C/S model, in which obtain arid agricultural articles by using web extraction technology,establish an ontology after a series of processing like segment and cluster, implementagricultural information vertical search engine, user interface and so on by combining opensource structure nutch and agricultural domain ontology. Software testing show this systemhas a good performance on robust and rationality.