排名數據卡方檢定的有限樣本性質

Finite sample property of chi-square test data for ranking

曾于芮(Yu-Jui-Tseng)

指導教授 ： 伍志祥

淡江大學/理學院/數學學系碩士在職專班/碩士(2018年)

https://doi.org/10.6846/TKU.2018.00702

摘要

排名資料在生活中常被關注是否具有相同的喜好度，在Anderson (1959)使用列連表卡方檢驗方式，Schalch (1979) 使用平衡不完全區塊設計，Jyh-Shyang Wu & Wen-Shuenn Deng (2017)則使用部分排序資料，以行的角度整理成一個 N×k 的列聯表，再用列連表卡方檢定統計量，建立一個檢定統計量。
本論文將延續Jyh-Shyang Wu & Wen-Shuenn Deng (2017)，推廣以列的角度整理成一個N×k 的列聯表，檢定統計量改為檢定統計量族，藉此方式說明喜好度並非以均勻機率的方式出現。

關鍵字

Anderson’s 檢定 ； 部分排序 ； 卡方檢定 ； 加強排名 ； 列連表

並列摘要

Ranking data is often concerned about whether it has the same preference in life. In Anderson (1959), the contingency table method is used, and Schalch (1979) uses the incomplete block design, Jyh-Shyang Wu & Wen-Shuenn Deng ( 2017) uses partial sorting data, sorts into an N×k contingency table from the perspective of the row, and then uses the collinear table to check the statistic to establish a statistic.
This paper will continue Jyh-Shyang Wu & Wen-Shuenn Deng (2017), promote the column into a N × k contingency table, the statistic is changed to the statistic family, which means that the preference is not Appear in a uniform chance.

並列關鍵字

Anderson’s test ； partially ranked data ； chi-square test ； imposed rank ； contingency table

參考文獻

Ranking data is often concerned about whether it has the same preference in life. In Anderson (1959), the contingency table method is used, and Schalch (1979) uses the incomplete block design, Jyh-Shyang Wu & Wen-Shuenn Deng ( 2017) uses partial sorting data, sorts into an N×k contingency table from the perspective of the row, and then uses the collinear table to check the statistic to establish a statistic.

This paper will continue Jyh-Shyang Wu & Wen-Shuenn Deng (2017), promote the column into a N × k contingency table, the statistic is changed to the statistic family, which means that the preference is not Appear in a uniform chance.