相似性を示す統計指標の数学的構造

• 山田 朋人

  東京大学生産技術研究所大学院課程

• 鼎 信次郎

  総合地球環境学研究所

• 沖 大幹

  東京大学生産技術研究所

書誌事項

タイトル別名

• MATHEMATICAL STRUCTURE OF A NEWLY-DERIVED STATISTICAL PARAMETER AS A SIMIIARITY INDEX

抄録

It is shown that a similarity parameter has new mathematical structure. Heretofore, the correlation coefficient is used for quantifying the correlation relationship between two ensemble members. Koster et al¹ introduced a statistical parameter, called Ω to quantify the similarity among several ensemble members with calculating the ensemble numbers and the two types of variances. However the mathematical structure of Ω had not been revealed in their studies. The present authors applied to derivate Ω for understanding the mathematical meaning of it. As results, we could have a knowledge that Ω consists of mainly two terms. One is the average value of cross correlation coefficients (ACCC) across all ensemble members. Another is the similarity of the mean value and the variance across all ensemble members. Therfore, the authorscan conclude that Ω shows the similarity of the ‘shape’ of all ensemble members and the mathematical characteristics is more capacious than the correlation coefficient. The paper ends with some remarks on the mathematical characteristics of ‘as a new evaluation methodology for the predictabity of numerical forecast.’ in monthly or seasonal time scale.

収録刊行物

• 水工学論文集

  水工学論文集 49 1-6, 2005

  公益社団法人 土木学会

キーワード

• Ω
• similarity parameter
• correlation coefficient
• AGCM
• ensemble forecast

詳細情報

• CRID: 1390282680148992256
• NII論文ID: 130003841995
• DOI: 10.2208/prohe.49.1
• ISSN: 18849172
• ISSN: 09167374
• データソース種別:
  • JaLC
  • Crossref
  • CiNii Articles
  • KAKEN
• 抄録ライセンスフラグ: 使用不可