Behaviormetrika
Online ISSN : 1349-6964
Print ISSN : 0385-7417
ISSN-L : 0385-7417
12 巻 , 17 号
選択された号の論文の4件中1~4を表示しています
  • Bradley Efron, Robert Tibshirani
    1985 年 12 巻 17 号 p. 1-35
    発行日: 1985年
    公開日: 2006/06/09
    ジャーナル フリー
    This is an invited review of bootstrap methods. It begins with an exposition of the bootstrap estimate of standard error for one-sample situations. Several examples, some involving quite complicated statistical procedures, are given. The bootstrap is then extended to other measures of statistical accuracy, like bias and prediction error, and to complicated data structures such as time series, censored data, and regression models. Several more examples are presented illustrating these ideas. The last third of the paper deals mainly with bootstrap confidence intervals. The paper ends with a FORTRAN program for bootstrap standard errors.
  • Chikio Hayashi, Mikio Yamamoto, Tsutomu Komazawa, Kinji Mizuno, Hiroyu ...
    1985 年 12 巻 17 号 p. 37-52
    発行日: 1985年
    公開日: 2006/06/09
    ジャーナル フリー
    In this study, a questionnaire was devised for comprehensive analysis of diverse social, cultural, and psychological factors that are considered to be strongly associated with coronary heart disease and stroke.
    The trial survey for the Japanese data was conducted, and the data were analyzed using Hayashi's Quantification Method 111. Questions were evaluated as to their usefulness in soliciting informative and relevant responses.
    Furthermore, the data were analyzed and correlated with physical examination findings, using Hayashi's Quantification Method 11. Analysis revealed the maximum latent root to be “Type A Behavior Pattern”. However, other factors, which seemed to be uniquely Japanese, also influence the incidence rate.
  • Ikuo Kabashima, Nozomu Matsubara
    1985 年 12 巻 17 号 p. 53-65
    発行日: 1985年
    公開日: 2006/06/09
    ジャーナル フリー
    Japanese Lower House elections are based on the medium constituency system, where by each electoral district is assigned a specific number of seats in the Diet according to its size. This makes it difficult to analyze election results across an entire district. To overcome this difficulty, we developed the MK ratio, which was obtained by dividing the number of votes cast for each candidate by the legal minimum vote. The paper then analyzed the thirty seventh general election dubbed as “The Tanaka Verdict Election” in comparison to the previous elections.
    The analysis revealed the following findings:
    1) Fluctuations in the number of successful LDP candidates are linked to movements in the floating vote. The floating vote consists of wavering LDP supporters who are attracted to or repelled by noted LDP candidates. This phenomenon is limited to urban electorates; 2) The Tanaka trial verdict did not affect all Tanaka faction candidates to the same degree. As Tanaka faction was conspicuously successful in the midst of the general LDP setback, gain within the faction were achieved by those politicians whose stance closely resembled the Tanaka style of regional advantage politics; 3) The opposition gains were the result of self-imposed losses on the LDP's part resulting from the alienation of the floating vote, and that this swing away from the LDP did not take the floating vote into the opposition camp; 4) High competition among LDP candidates in the same district did not result in opposition gain. The implications of the results of the recent election for the contemporary Japanese political system were also discussed.
  • Yukio Yanagisawa
    1985 年 12 巻 17 号 p. 67-79
    発行日: 1985年
    公開日: 2006/06/09
    ジャーナル フリー
    Combinatorial cluster analysis includes the single linkage, complete linkage, and other methods, all of which arrange individuals as a dendrogram. Consider the application of any of these methods to a set of n individuals, which are represented by points in m-dimensional Euclidean space Rm. Suppose that the positions of n-1 individuals are fixed, but that one is allowed to vary. Then the form of the resulting dendrogram depends on the position of the n-th individual. We show that the different possible forms divide Rm into domains; within each domain, the lengths assigned to individuals vary, and therefore the clusters will differ. The domains are bounded by m-dimensional spheres and (m-1)-dimensional hyperplanes.
feedback
Top