| Great philosophers all pay much attention to the study of history. Francis Bacon once said " histories makemenwise". And "gaining new knowledge throughreviewing what has been learned " has long been the axiom of the Chinese people. It is no doubt that historical research is important. This doctoral dissertation is concerned with the early d evelopment o f t he i dea of s tatistical i nference. It introduces and explores i n detail the early academic document of the three schools of statistics, in which the author's systematic treatment of the idea of inverse probability and likelihood can be seen as the most recent research development of its kind in China. Discussions about the essence of statistical thoughts are also presented in the light of philosophy. This leads the author to come to the conclusion that the so-called statistical thoughts are nothing but the combination of metaphysical beliefs and atomic facts through which statistical ways of thinking and treatment could be developed. Also, the author discusses the inferential logic under uncertainty, and points out that the fatal weakness of the traditional frequency school is that the way of how to produce a proper null hypothesis has never been clearly stated. From the viewpoint of how to draw inference in uncertain circumstances, this paper touches the history of statistical science and makes a few reflections on the existing problems in contemporary statistics.This paper consists of five chapters. The basic contents" are as follows. Chapter One: IntroductionIn this chapter, the author first discusses the importance and the practical meaning of historical research in statistics, he points out that there exists an unfavorable phenomenon in the development of statistics that statistics has been highly armed with mathematics and yet it is really lack of genuine statistical thinking. A possible remedy for correcting it lies in the historical research of statistics. Second, the framework and breakthroughs of this paper are presented in which several points may deserve more future attention. Third, both domestic and foreign historical research issues of statistics are mentioned.Chapter Two: The emergence of the concept of probabilityThis chapter contributes mainly to the early forming stage of the concept of probability, namely, (1) the correspondence between the two great mathematicians Pascal and Fermat in 1654. Not only did the correspondence solve the "division problem", but also it made the two men founders of probability theory. (2) Huygens printed his famous book "Of the laws of chance" in 1657. This book provided an opportunity for mathematicians at that time to take up the course, so that they can well understand the propositions, questions, explanations and the generalization of probability. (3) The law of large number was demonstrated in James Bernoulli's "The art of conjecturing", which was published after his death in 1713. This was a great event, which signified the end-up of the formation of probability concept and the beginning of mathematical probability development.Chapter Three: The early forming stage of the concept of inverse probability This chapter introduces in detail three articles that cast great influence on the early development of inverse probability (i.e. the early history of Bayesian School). They are (1) "An essay towards solving a problem in the doctrine of chance" published in 1764 after Bayes' d eath. B arnard, a n o utstanding s tatistician s aid t hat t his p aper i s definitely one of the most important papers in the history of science. (2) Laplace printed in 1774 "Memoir on the probability of the causes of events" .This can be seen as a milestone. It was this paper that made Bayesian School become known in the field of mathematics, and moreover we can still learn something new from it at the present time. (3) The error distribution was derived in Gauss' "Theory of the Motion of the Heavenly Bodies Moving about the Sun in Conic Sections" which was published in 1809. Again Gauss also offered in this book the famous least square...
|
PDF Full Text Request