東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Data preparation for clinical data m...

Lin, Jau-Huei.

FindBook

Google Book

Amazon

博客來

Data preparation for clinical data mining in developing a problem list proposing system.

紀錄類型:	書目-語言資料,印刷品 : Monograph/item
正題名/作者:	Data preparation for clinical data mining in developing a problem list proposing system./
作者:	Lin, Jau-Huei.
面頁冊數:	153 p.
附註:	Source: Dissertation Abstracts International, Volume: 69-02, Section: B, page: 0766.
Contained By:	Dissertation Abstracts International69-02B.
標題:	Biology, Bioinformatics. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3302493
ISBN:	9780549481508

Data preparation for clinical data mining in developing a problem list proposing system.
Lin, Jau-Huei.

Data preparation for clinical data mining in developing a problem list proposing system. - 153 p.

Source: Dissertation Abstracts International, Volume: 69-02, Section: B, page: 0766.

Thesis (Ph.D.)--The University of Utah, 2008.

This dissertation was focused on the process of preparing clinical data from its raw format in an electronic health record system to a format appropriate for data mining. The purpose of this data mining effort was to extract logic for a decision support system able to propose medical problems to clinicians. In order to develop a process that could be applied to a large-scaled development, i.e., a decision support system capable of detecting hundreds of medical problems from thousands of variables available in the database, the data preparation process should be systematic and automatic.

ISBN: 9780549481508Subjects--Topical Terms:

1018415
Biology, Bioinformatics.

Data preparation for clinical data mining in developing a problem list proposing system.
LDR:03286nam 2200289 a 45 001 963668
005 20110831
008 110831s2008 ||||||||||||||||| ||eng d
020 $a 9780549481508
035 $a (UMI)AAI3302493
035 $a AAI3302493
040 $a UMI $c UMI
100 1 $a Lin, Jau-Huei. $3 1286731
245 1 0 $a Data preparation for clinical data mining in developing a problem list proposing system.
300 $a 153 p.
500 $a Source: Dissertation Abstracts International, Volume: 69-02, Section: B, page: 0766.
502 $a Thesis (Ph.D.)--The University of Utah, 2008.
520 $a This dissertation was focused on the process of preparing clinical data from its raw format in an electronic health record system to a format appropriate for data mining. The purpose of this data mining effort was to extract logic for a decision support system able to propose medical problems to clinicians. In order to develop a process that could be applied to a large-scaled development, i.e., a decision support system capable of detecting hundreds of medical problems from thousands of variables available in the database, the data preparation process should be systematic and automatic.
520 $a In this study, a model called "data preparation framework" was developed for the process of data preparation. In the proposed model, data are first retrieved from the data source and transformed into a flattened table format. Then three preprocessing treatments---missingness representation, general statistical heuristics, and semantic filtering---are applied to rectify certain data quality problems inherited from the data source. Automatic variable selection algorithms are then used to reduce the number of variables according to their relevance for detecting medical problems. Clinical expertise is used to judge the clinical relevance of each variable in the reduced variable subset, which has been chosen automatically. After expert verification, a data set suitable for further machine learning can be created.
520 $a To validate the proposed model, a number of experiments were conducted. Two types of measurements were assessed as indicators of performance: clinical relevance of selected variables, judged by clinicians; machine learning performance, evaluated using the area under the receiver operating characteristic curve. The results show that each of the three introduced preprocessing treatments significantly improved the performance, indicated by both types of measurements. In addition, the data sets prepared using the proposed process achieved high clinical relevance and relatively good machine learning performance, compared with various data preparatory pathways tested in the experiment.
520 $a The proposed data preparation framework incorporates automatic methods with heuristic preprocessing treatments for the potential challenges within a large-scaled development. The automatic methods and heuristic rules can help reduce the demand on manual work in preparing appropriate data and thus make the development feasible.
590 $a School code: 0240.
650 4 $a Biology, Bioinformatics. $3 1018415
650 4 $a Computer Science. $3 626642
690 $a 0715
690 $a 0984
710 2 $a The University of Utah. $3 1017410
773 0 $t Dissertation Abstracts International $g 69-02B.
790 $a 0240
791 $a Ph.D.
792 $a 2008
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3302493