Abstract
For an adaptation to the new frame of diagnostic theory of handwriting proposed in the companion paper, we adopt Stability and Specificity of a diagnostic variable instead of Constancy and Rareness of a diagnostic feature(Toya, 1965). Then we prescribe a diagnostic value to a diagnostic variable with a weighting function of Stability and Specificity on an evaluation plane. Based on these new implementations for diagnostic theory of handwriting and a wide survey of Japanese hands, we formulate the theory of writer discrimination of a written document as a statistical theory of hypothesis test. We present three methods of writer discrimination:(1)testing hypothesis with one variable(Ohshio, 1975), (2)comparing similarity by means of Euclidean distance between two points in a diagnostic space(Ohshio, 1978), (3)comparing skillfulness of handwriting by means of the developmental law. We applied these methods to the issue of writer discrimination in the Sayama case.
