2004Ouyou toukeigakuOpen access

On Avoidance of the Over-fitting in the B-Spline Non-parametric Regression Model

Hirokazu Yanagihara, Megu Ohtaki

Open full text 2 citations

Abstract

複雑なトレンドを持つデータに対して平滑化をおこなう場合,B-スプラインノンパラメトリック回帰モデルを用いた手法は,単純でかつ短い計算時間で結果を得ることができるといった利点を持つ.しかしながら,このモデルは本来柔軟なモデルであるため,適切に用いない場合に,柔らか過ぎる適合,過剰な適合を起こす場合がある.本論文では,基底関数の個数と平滑化パラメータを決定するための情報量規準を改良することでこの過剰適合を回避する手法を提案する。

Open-access reader

About this research paper

What this paper is about

複雑なトレンドを持つデータに対して平滑化をおこなう場合,B-スプラインノンパラメトリック回帰モデルを用いた手法は,単純でかつ短い計算時間で結果を得ることができるといった利点を持つ.しかしながら,このモデルは本来柔軟なモデルであるため,適切に用いない場合に,柔らか過ぎる適合,過剰な適合を起こす場合がある.本論文では,基底関数の個数と平滑化パラメータを決定するための情報量規準を改良することでこの過剰適合を回避する手法を提案する。

Why it matters

OpenAlex reports 2 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

複雑なトレンドを持つデータに対して平滑化をおこなう場合,B-スプラインノンパラメトリック回帰モデルを用いた手法は,単純でかつ短い計算時間で結果を得ることができるといった利点を持つ.しかしながら,このモデルは本来柔軟なモデルであるため,適切に用いない場合に,柔らか過ぎる適合,過剰な適合を起こす場合がある.本論文では,基底関数の個数と平滑化パラメータを決定するための情報量規準を改良することでこの過剰適合を回避する手法を提案する。

Key concepts: Mathematics, Parametric statistics, Regression analysis, Regression, Statistics, Spline (mechanical), Nonparametric regression, B-spline

Related papers

Back to paper searchBrowse research topicsOriginal source
On Avoidance of the Over-fitting in the B-Spline Non-parametric Regression Model — Research Paper | ScholarLens