| Summary: | 碩士 === 輔仁大學 === 資訊管理學系 === 98 === In recent years, because of the number of small family increases, the travel trends have changed, and the government vigorously promotes the homestay industry, so that the number of homestay has increased year by year, but compared to the standard hotel-style services, homestay industry is approachable and has a friendly way to get along for ease of modern life so that people can pursue a higher degree of leisure life. As the shortage of information of accommodation sources, information is not unilaterally offered by the homestay industry but is dependent on tourists who wrote online articles to deliver the information. How to filter out the information to represent homestay industry would be a worthwhile research issues.
In this study, latent semantic analysis (LSA) and singular value decomposition (SVD) are used to extract keywords to classify the homestay articles as well as make the tagging for articles. In such a way users can find keywords easier for classification of the areas of concern. In this study, the keyword extraction methods are used to identify keyword models that are close to social tagging in order to give appropriate recommended taggings.
In this study, if the same words appear in different articles more often, they will have higher correlation values in the articles they do not appear. Compare the keywords extracted by LSA, and those extracted by TF-IDF weighted method, the precision and recall values of LSA are better than those of TF-IDF so that LSA is more suitable to judge the keywords of artilcles. In this study, the precision and recall vlues obtained are up to a certain accuracy level so that it helps articles to establish close social homestay tags of the keywords. With the results, it shows that users can search information and references easier with the help of the methods applied in this study.
|
|---|
