The Application Research Of Data Mining In GUILIN Tourism Information

How much does Guilin benefit from so many tourists? Guilin tourism achieved historic new achievements in 2005: there were 12.0508 million tourists in the city, increased 8.43% for the same period, and there were 1.0009 million inbound tourists, increased 23.62% for the same period. There were 11.0499 millions domestic 11.0499 millions, increased 7.21% for the same period. The gap is very obvious, from the pillars of touring benefit Guilin social economy. GDP of Guilin in 2005 was 53.67 billion Yuan, and total touring revenue was 5.795 billion Yuan. and increased 15.57%,but contribution rate was only about 11%.If contrast according to adding value, tourism occupies a much small proportion, no more than 6%.On the one hand, there are so many tourist, on the other hand touring revenue is very low. Where is the problem on earth? There will be a feasible mean, look for keys from touring information data using Data Mining technology.Data Mining means picking up hidden, useful information and knowledge process from vast, incomplete, noisy, vague and random data.Its expressive forms are concepts, rules, patterns, and so on. Data mining function includes finding concept descriptions, association rules, classification and predicition, clustering, trend analysis, deviation analysis, similarity analysis. Among them, associated regulations, classification and prediction,and cluster analysis are most frequently applied in tourist information. So this paper lays emphasis on study of data mining process and classification data mining technology in theory part.The paper is basis of the classification technology of data mining, using SAS/EM data mining tools,data of touring investigation of Guilin City Touring Agency in 2005, perform classification mining from two aspects.That is factor of influencing tourists' consumption, factor of influencing tourists' general assessment to Guilin touring. In this process, realize classification data mining's whole process perfectly, includes: confirming data sources, mining targets, data pretreatment, using SAS/EM tool, creating decision tree, obtaining corresponding regulations, analyzing results. Among these, the data pretreatment use x~2 statisti cal testing to select the properties associated with the mining targets. This step is very important, because the algorithm c4.5 that SAS/EM decision tree node supports, which requires too much for effectiveness of property. Sum up and prospect, according to research status at last.