Design And Implementation Of A Tourism-oriented Vertical Search Engine

With the rapid development of the level of people’s living, tourism becomes an important way of people’s entertainment. However, before the bourney, making a perfect journey plan confuses many people, including where and how to go, where to live and so on. Internet is an important information source for tourists. However, there are too many web pages in internet, and the traditional search engine would return too much unrelated information with tourism for the users.Under this background, this dissertation foucs on the design the implementation of the tourism oriented vertical search engine. Comparing with the normal search engine, the vertical search engine will provide more specialized, more accurate and much deeper information for the tourists. That is because the system is structured based on domain data.The works of this dissertation are as follew.Firstly, we survey the requirement of system, including data requirement and function requirement. The data requirement of tourism oriented vertical search engine determines the integrating information boundary of the system. It analyses the type of data which the tourist fouce on, such as the introduction of tourist attractions, hotels, traffic. While the function requirement analysis would model the business process of vertical search engine, which is the fundamental of function design.Then, based on the requirement analysis, we degine the whole architecture of the tourism oriented vertical search engine Vertical search engine is a system with complex inner compositions, including foreground and background. The foreground department is the application program interface(API) for the users’ queries, while, web page integration is complicated in the background. Many functions, such as crawler, page filter, information retrieval are in the background. Background and foreground run in parallel.In the following dapartement, we focus on the introduction of the design and implementation of 3 kernal functions. In this dissertation, we design the crawler based on breadth-first search to download the whole web pages including text, images and videos. The web page filter is designed based on Navie Bayes classifier, according topic dictionary. While Lucene provides text information retrieval function, we also design the meta-data retrieval function for the pages.At the end part of the dissertation, we explain the realization tourism oriented vertical search engine. Besides, we carry out an experiment for the system to verify our design.