ANALYSIS OF ERRORS IN CLASSIFICATION

スポンサーリンク

概要

• 論文の詳細を見る

• Frequency data obtained by classification may be biased because of imperfect reliability of method or subjects. A model for classification is proposed and formulae for estimating possible biases are developed by means of a segment model. In this model, the following assumptions are adopted; (1) the psychological continuum is divided into several successive segments, (2) the events to be classified distribute equally over the segments, (3) the criterion of classification is also equally distributed, and (4) the number of related segments and their relationship are the determing factors of classification. The analysis is expanded to the case of continuous distribution and a theoretical formula and a practical equation are suggested to calculate the expected frequency.

• 公益社団法人 日本心理学会の論文

著者

• Tanaka Yoshihisa

  University Of Tokyo

関連論文

• Generalized Methods for the Construction of Psycho-physical Scale by the Method of Fractionation and Multiplication
• ANALYSIS OF FACTORS THAT DETERMINE SUBJECTIVE WEIGHT
• GENERAL PRINCIPLES FOR THE CONSTRUCTION OF ORDINAL SCALE:1. PRINCIPLES OF FREQUENCY DISTRIBUTION
• ANALYSIS OF ERRORS IN CLASSIFICATION

スポンサーリンク