Recently, various kinds of learning models have been built and discussed in psychology and control engineering. Learning concept is vital to represent a social decision process. There is, however, no learning model for this process due to its difficulty. This paper attempts to throw light upon this difficulty by giving some learning points which the model should have, and construct a learning model for social decision process which involves recognition and decision making models as sub-models.This paper call this model as“ Adaptive learning model”. Lastly, we prepare two these models as two major parties in the political system, and simulate these decision makings using two different learning policies so as to evaluate adaptive learning model.
Confining the scope to the citizens' requst data in the questionnaire to all households conducted by Kobe City Office, this paper reports the analytical results of the following two subjects. Firstly, differences of the requests in region, sex and age group and the kinships between the requests are presented. Secondly, an behavioral approach to estimate an importance-weight of each request and its results are explained. The largeness of sample numbers of the based data permits us to analyze them in detail and obtain unique results.
One hundred and thirty one medical students(111males and20females)were requested to reply to the MMPI questions(Japanese version)during and56days after the entrance examination. The greatest difference between the two tests seemed to be that the tension influence was not prevalent at the second test. A cross-table analysis revealed that the items which resulted in the same answers in the two tests are concerned with the F, Sc or Pa scales, whereas the items which led to the different answers are concerned with the Ma scale. By means of factor analyses of both tests based on the validity and clinical scales, the same three factors were extracted: such as self-criticism with irritability, suspicion and aggression, introversive unsociability. The above results showed that the MMPI would be recognized as well defined as the character trait personality inventory.
Known social decision methods are the majority rule, the unanimity rule, and the dictatorial consulation method. In this paper we introduce two characteristic measures, i. e. dictatorial rate γ and biassing rateδ, for treating social decision methods. The majority rule and the unanimity rule have0D-rate. But the dictatorial consultation method has1D-rate. The simple majority rule and the dicatorial consultation method have0B-rate. And the unanimity rule has1B-rate. There are many decision methods, whenever their rates satisfy0≤γ≤ 1, |δ|≤ 1-γ. The majority rule is sufficient for legislative organs. But for administrative organs neither the majority rule nor the dictatorial consultation method are appropriate. So we propose decision methods the rates of which satisfy0.25≤γ0.5, δ=0.For example if we set γ=0.5, δ=0, a bill will pass when the approbation is more than three quarters, that will be rejected when the opposition is more than three quarters, and otherwise the boss can decide as he likes.