逆回帰問題における高精度な推定量の開発に関する研究

逆回归问题高精度估计器的开发研究

• 批准号: 14780166
• 负责人: 井上 淳
• 金额: $ 0.83万
• 依托单位: Waseda University
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Young Scientists (B)
• 财政年份: 2002
• 资助国家: 日本
• 起止时间: 2002 至 2003
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-14780166/
• 关键词: 線形校正; 逆回帰

本年度の研究計画に従い,昨年度に行なったシミュレーションの結果を分析し,考察を加えた後に,数学的に詳細な議論を行なった.具体的には,先ず,Inoue(2003)で得られている一連の結果(不均一な誤差分布・分散下での回帰ベクトルの推定量の構築とその性質に関する一連の結果),および昨年度に実施したシミュレーションの結果を併用しつつ,申請者が考案した複数の推定量族の挙動を数学的に厳密に調べていった(※数学的な議論の道筋を違えないために,小規模なシミュレーションも再度必要に応じて実行した).その結果,異質な条件下で得られたサンプル群の情報を結合して回帰の逆推定を行なう際,従来の推定方法には問題点が多数観察されることを,導出した精度の評価式にもとづいて数学的に(明示的に)指摘することができた.これは従来知られていない新しい結果である.次に,ある判断基準の下,提案した推定量族の中で最適なものを見つけ出すことができた.続いて,Inoue(2003)で考案されている反復的な推定手法も活用しつつ,更なる推定量の改良に取り組んだ.結果的に,先述の推定量族内で最適な推定量よりも精度の良いものを導出することができた.推定量族の小標本的な挙動については,限定的な結果が得られるにとどまった.しかし,大標本的挙動については一連の成果が得られた(※なお,小標本的な挙動については,研究期間終了後も引き続き独自に研究を継続していく).現在は,一連の研究成果を取り纏めている.終了次第,査読付き学術論文誌に研究成果を投稿し,公表する.

In this year's research program, we will conduct an analysis of the results of this year's research program, which was conducted last year. We will examine the mathematical theory and practice of mathematics after the addition of money. For specific information, first, Inoue (2003), we got the results of the first round of experiments (the results of the experiments were obtained yesterday. The results of the results show that the results of the results are different. Applicants are required to calculate the number of cases in the examination, calculate the number of numbers, and calculate the number of mathematical problems in mathematics. (in the discussion of mathematics, it is necessary to do so again). The results show that, under the conditions, the results show that the combination of the statistical data and the inverse presumption of the presumption method is effective, and that the majority of the questions about the presumption method are not valid. This leads to the accuracy, accuracy, You can learn more about the new results. For the second time, under the standard of judgment, the proposal is to determine the number of people in the quantitative family. Inoue (2003). The presumption technique is used flexibly, and the presumption quantity is improved. The results show that the accuracy of the most accurate quantitative analysis in the family is accurate. It is presumed that the activities of the small headers of the quantity family are not valid, and the limited results are satisfactory. The activities of the big headers are successful, and the results of the big headers are successful. (after the study period, after the study period, you can do research alone.) At present, the research results are drawn from the results of the research. For the first time, please submit contributions to the research results of academic journals and journals, and submit a public form.