Dr. Tang Man-lai, Associate Professor of the Department of Mathematics, won the 2nd Class Prize in the Natural Science of Yunnan Province award scheme in 2011 for a collaborative research project entitled “Biostatistics: Theory and Methods”. The project examined the insufficiencies of statistical methods that are currently applied in biological medicine, clinical research and pharmacy.

 

Dr. Tang and his collaborators from Yunnan University successfully investigated how to improve the defects of existing statistical methods by suggesting a more accurate method to handle and analyse incomplete discrete data in a more effective manner, thereby enhancing the accuracy and representativeness in statistical results.

Members of the research team from Yunnan University were: Professors Tang Niansheng, Dean of Faculty of Mathematics and Statistics, Wang Xueren, former Vice-Chancellor and President, Li Huiqiong, and Chen Xuedong.

Dr. Tang said: “In analysing collected data, we often find it difficult to handle incomplete discrete data. Incomplete or missing data refer to information that cannot be collected from the samples, such as sensitive data like income and sexual orientation. In some cases, the selected samples may fail to provide necessary information in follow-up clinical research due to adverse effects caused by a treatment. Unlike continuous data, discrete data can take on only a finite or countable number of values (i.e. integer values). Therefore, special statistical procedures are desired and required for handling incomplete discrete data.” Ignorance of incomplete discrete data in any research will significantly reduce the credibility and representativeness of the conclusion and final outputs.

浸大數學系副教授鄧文禮博士憑一項與雲南大學的協作研究項目「生物統計的理論和方法研究」，就現時應用於生物醫學、臨床測試、藥劑學等範疇的統計方法作出改善方案，獲得2011年度雲南省自然科學獎（二等獎）。

 

鄧博士和研究團隊為改善現有統計方法的缺點，研究如何有效處理和分析缺失離散數據，設計出更精確的統計方法，從而提高統計結果的精確度和代表性。

 

項目協作者包括雲南大學的數學及統計學院院長唐年勝教授、前任校長王學仁教授、李會瓊教授和陳雪東教授。

 

鄧博士說：「收集統計數據時，我們不時遇到缺失數據和離散數據的難題。缺失數據一般指收集數據時被訪者未能提供的資料，例如薪金、性傾向等敏感數據，或病人接受某項治療後出現副作用需停止治療，因而未能完成臨床測試所需的資料。與連續數據不同，離散數據只能在有限或可數的數值上取值，通常以整數單位計算。當所收集資料涉及缺失或離散數據，我們需要應用特別的統計方法分析。」統計項目如出現大量缺失或離散數據，將直接減低統計結果的可信性和代表性。