LER Data Mining Pilot Study Final Report

Young, Jonathan ; Zentner, Michael D. ; McQuerry, Dennis L.

Pacific Northwest National Laboratory (U.S.)

关键词: Information Systems;    Classification;    Document Types Data Mining;    Information Retrieval;    99 General And Miscellaneous//Mathematics, Computing, And Information Science;   

DOI  :  10.2172/15020763
RP-ID  :  PNNL-14910
RP-ID  :  AC05-76RL01830
RP-ID  :  15020763

美国|英语

来源: UNT Digital Library

【 摘 要 】

LERs consist of a one page standard form with a standard header and free text data, followed by additional continuation pages of free text data. Currently this LER data is analyzed by first inputting the heading and text data manually into a categorical relational database. The data is then evaluated by enumeration of data in various categories and supplemented by review of individual LERs. This is labor intensive and makes it difficult to relate specific descriptive text to enumerated results. State of the art data mining and visualization technology exists that can eliminate the need for manual categorization, maintain the text relationships within each report, produce the same enumerated results currently available, and provide a tool to support potentially useful additional analysis of the informational content of LERs in a more timely and cost effective manner.