A multivariate analysis of risk factors for cerebrovascular disease (CVD) has been applied to 78 CVD cases in 1, 419 residents during a six-year follow-up period in Hisayama, Kyushu Island, Japan. Age or systolic blood pressure among thirteen variates showed the largest coefficient in the discriminant function in standard unit in both CVD as a whole and cerebral thrombosis for both sexes. After the probability (P) of developing disease was calculated by a logistic function using the estimated coefficients, about one-third of the subjects were assigned to the highest decile group of P and the subsequent four-year follow-up study showed that over eighty per cent of the newly developed 22 cerebral thrombosis occurred in the highest decile group. Increasing number of the variates did not necessarily improve the correct prediction rate for cerebral thrombosis and six or seven variates were most predictive in the present analysis.
Two models concerning with choice probability were reviewed: one by Luce and the other by Tversky. Traditionally, application of this type of choice models has been limited to choices repeated by respective individuals. Hence, discussion was focussed upon the basic postulates of the models and the applicability to pooled data over a population of individuals. With regard to the Luce's model, it was pointed out that the scale obtainable from the model that successfully predicts the choice probabilities can not be the variable representing magnitude of sensory intensity or of preference. With regard to the Tversky's model, its strategic implications in campaign were discussed in detail and an empirical study was referred to.
A computer program called POET generated “poems” through a system which incorporated (1) transitional dependencies of words, (2) transformational structures of language, and (3) semantic, or imagery, aspects in the decision about a general theme of a poem. It was assumed that each sentence in a “poem” contains a single “thematic word” as the main subject. Thematic words were selected on the basis of the similarities of words to the major thematic word (which was to be given by the experimenter) in an image space defined according to Osgood's Semantic Differential Space. The other words in the individual sentences were assumed to be generated on the basis of both transition probabilities of word associations under the given theme and choice probabilities of transformational rules in language habit. Some of the “poems” generated by the system were introduced and discussed in relations to psychological processes of sentence generation.
Recently, a large number of techniques of numerical taxonomy or cluster analysis are proposed and studied by many workers, and the needs for these techniques arise in many fields of applied science. Especially, the method of hierarchical cluster analysis is used widely, since all of these methods are suitable for various types of data and can be simply carried out. However, though the investigations for evaluation or comparison of these method are only little discussed in formal works, it is quite important and necessary to discuss them in order to study cluster analysis. Furthermore, we propose the fuzzy distance which is a new index of evaluating and comparing relationship between two relation matrices, the original similarity matrix S and S* which is derived from S by executing hierarchical clustering algorithms. Finally, numerical examples for several artificial datas are investigated by four well-known clustering methods.