陳宏2006-07-262018-06-282006-07-262018-06-282003-07-31http://ntur.lib.ntu.edu.tw//handle/246246/20980不完整數據的統計分析，於近年來極受重視。我們考慮使用Cronbach 量測測驗 中含闕失數據時對信度所造成的影響;另在cDNA 微陣列所測量的數據，會因 染劑、點印基因探針針頭而造成系統性誤差，我們對常用的幾種正規化方法進行 比較，並討論這幾種方法間的共同性及差異，並提出一個正規化的新方方法。In recent years, statistical analysis for handling incomplete data has drawn a lot of attentions. In this project, we consider the meaning of Cronbach  when the test includes missing responses. We also consider how to do normalization with cDNA microarray data. It is known that both the efficiency of dye and the gene probes print-tips will introduce systematic bias into the observed measurements. We do a comparison of commonly used normalization methods. In addition, we also propose an alternative normalization procedure.application/pdf400167 bytesapplication/pdfzh-TW國立臺灣大學數學系暨研究所不完整數據電腦適性測驗Cronbach α係數正規化cDNA 微陣列Incomplete dataComputerized adaptive testingItem response theoryCronbach αNormalizationcDNA microarrayreporthttp://ntur.lib.ntu.edu.tw/bitstream/246246/20980/1/912118M002005.pdf