| Summary: | 碩士 === 南台科技大學 === 資訊管理系 === 91 === In this thesis, we use two mining methods and two ways to discover the most adaptive readers of one book, and adaptive books of readers. With cluster method. We use cluster to find the readers whose best fit the special book. We use records of readers in the database of library, and according age, academic and department to criterion which cluster is adaptive for one reader. Then, we appoint a book, and research a cluster which book best fit. Finally, find the readers whose never borrow the book which we appointed. The readers are most adaptive readers of one book. Another way. We use the same method to find the most adaptive books of one reader. Like before process, base on data of readers. We clustering all reader to several clusters. Then, research the books of a special cluster, and find the readers which haven’t the record of those books in this cluster. We can identify those books are the those reader need.
In the other method, we use sequential pattern method to discover the most adaptive readers of one book, and discover the readers’ most fit books. First, base on records in the library’s database. The readers’ number is primary sort key, and the records’ date is slavery sort key. We can find a reader’s reading sequence. Then, according the minimal support to judge all itemsets which is combination by reader’s borrowed books. Finally, delete the subsequence of maximal sequence itemsets which had satisfying minimal support. If a book which we appointed is in those maximal sequence itemsets. We can use the borrowing sequence to research the best fit reader. Another way, we user the same method and process to discover the adaptive books of readers.
|
|---|
