| Budget Amount *help |
¥3,000,000 (Direct Cost: ¥3,000,000)
Fiscal Year 1999: ¥600,000 (Direct Cost: ¥600,000)
Fiscal Year 1998: ¥500,000 (Direct Cost: ¥500,000)
Fiscal Year 1997: ¥1,900,000 (Direct Cost: ¥1,900,000)
|
|---|
| Research Abstract |
This study has provided us with the software and the hardware for rating psychological impressions to noise stimulus.
1.Concretely, the impressions caused by the presented stimulus is inputted by the push buttons which are assigned by seven categorized impressions : F1(very calm), F2(quite calm), F3(slightly calm), F4(medium), F5(slightly noisy), F6(quite noisy), F7(very noisy). On the other hand, the input noise L is logically rated by the membership function μT(L) which implies the grade of the truth of the proposition T : "L is noisy on the basis of calmness. The value of the proposition T is modified by the linguistic hedge τi which implies very, quite, ..., and L is ultimately rated by the expression of the truth qualification τi(μT(L)). The validity and the usefulness of the rating model which we call the method of fuzzy successive categories, have been confirmed from the following viewpoints : (1)μT(L) is significantly influenced by the presented noise since its universe of disco
… More
urse is directly the noise level. (2)τi(s) is hardly influenced by the noise since its universe is the logical domain. Practically, μT(L) and τi(s) are specified by the S-type and the triangular type respectively.
2.The present fuzzy method is further developed into a dynamic one in order to predict the output of psychological quantities to the input stimulus by use of the fuzzy reasoning. This dynamical one includes in itself the function to estimate μT(L) by use of the past data of the pair of the input and the output. This dynamic one is also useful to estimate the level L1, L0.5 and L0 respectively which satisfy the following condition : μT(L)=1, μT(L)=0.5 and μT(L)=0. Their levels show the standard of subjective judgment and the changes of the levels by context effects is clearly obtained although the conventional method seems to be impossible.
3.A method of estimating the membership functions μFi(L)=τi(μT(L))(I=1, ..., 7) is also studied. Concretely, μFi(L) is specified by the trapezoidal one in advance and its parameters are estimated by use of a modified maximum likelihood method. As the result, we can see the psychological judgment including the uncertainty which consists of the fuzziness and the randomness. Less
|
