書誌事項
- タイトル別名
-
- Projection Index with Empirical Distribution Function
- ケイケン ブンプ カンスウ オ モチイタ アラタ ナ シャエイ シヒョウ ノ テイアン
この論文をさがす
抄録
In this paper, we propose a new method of projection pursuit which extends a concept of 'interestingness' and develop a projection index for it.<BR>It is generally difficult to see the structure of data when the dimension of data is high. Therefore, many studies have been reported on the methods which reduce high dimensional data into a smaller number of dimensions. Among those methods, projection pursuit was developed by Friedman and Tukey (1974) in order to search for an 'interesting' linear projection of multidimensional data. The feature 'interestingness' is defined as the difference between a distribution of target data and normal distribution. To measure this feature, many projection indices have been proposed to formulate the 'interestingness' mathematically.<BR>However, measuring the difference from normal distribution by such projection indices does not always give 'interestingness' because 'interesting' structure depends on data or on purposes of the analysis.<BR>Thus, we develop a method with a new projection index that a user can pre-define an 'interesting' structure according to the feature of the data or the purpose of the analysis. Our index measures a distance between the distribution of the projection of target data and that of the projection of pre-defined 'uninteresting' data. We show the effectiveness of the proposed method with numerical examples and actual data.
収録刊行物
-
- 応用統計学
-
応用統計学 32 (1), 17-29, 2003
応用統計学会
- Tweet
詳細情報 詳細情報について
-
- CRID
- 1390001204441455232
-
- NII論文ID
- 10011680028
-
- NII書誌ID
- AN00330942
-
- ISSN
- 18838081
- 02850370
-
- NDL書誌ID
- 6701741
-
- 本文言語コード
- ja
-
- データソース種別
-
- JaLC
- NDL
- Crossref
- CiNii Articles
-
- 抄録ライセンスフラグ
- 使用不可