In our creative activities, it is important for us to interpret information based on various viewpoints. Previous interfaces for information visualization do not display varying information states but display a single state only. Therefore, it is difficult for us to notice and interpret information differences with previous interfaces. In this paper, we propose an animation interface to interpret preference differences. Users express their own preference by arranging keywords, and their arrangements are compared on this interface using animation. By some experimental results, we verified that the proposed interface can support preference difference interpretation.
The paper proposes a method that analyzes discussions in threads of bulletin board sites. The method describes interests of users based on events in order to deal with the ambiguity related to the interests. Here, the events mean that specific objects, actions, impressions, and so on occur in the discussions. The paper mainly discusses two extraction methods. One method extracts important discussions from threads based on two kinds of events specified by the interests. The other method extracts characteristic expressions from the important threads by getting rid of the expressions incompatible with the interests. Here, the expressions are composed of words related to specific rows of parts of speech. The paper verifes the effect of the extraction methods by applying them to 21,867 articles collected from bulletin board sites. The paper also shows that the extraction methods can assign higher orders to important threads and they can extract characteristic expressions compatible to the subjectivity of the users through the discussion of the experimental results.
We propose an identification method of f in Choquet integral (C) ∫ fdμ using pairwise comparision. In many cases, the comparison words used by AHP are assigned fixed comparison values. Our experiments show that the correspondence table's parameter among the comparisons words and the comparison values change by the comparison objects. Input values of Choquet integral should have same evaluation unit among all evaluation criteria. In our methods, examinees reply not only pairwise comparisons, but also ideal and satisfaction points. By varying the correspondence parameter, all satisfaction and ideal point's values have the same values in the evaluation criteria. Lastly, we show some numerical examples that are domain sets are finite discrete set and continuous values sets.
The present study considers a fuzzy color system in which three membership functions are constructed on the RGB color triangle. This system can process a fuzzy input to an RGB system and output the center of gravity of three weights associated with respective grades. Three membership functions are applied to the RGB color triangle relationship. By treating three membership functions of redness, greenness, and blueness on the RGB color triangle, a target color can be easily obtained as the center of gravity of the output fuzzy set. In the present paper, the differences among fuzzy input, inference output, and chromaticity are described, and the relationship between inference outputs for crisp inputs and inference outputs for fuzzy inputs are shown on the chromaticity diagram.
Subjective uncertainty at the bottom of various evaluations is the most essential subject that was the reason fuzzy theory was proposed. Related research areas are widely spread such as decision making, data analysis, information retrieval, psychology, and human computer interaction and so on. As a basis for the various research areas a mathematical interpretation of evaluation occupies an important position. In fuzzy theory, possibility and necessity evaluations have been utilized as the ways for evaluating a vague achievement in an ambiguous situation. However, mathematical interpretations for the evaluations were insufficient. In this article, possibility and necessity evaluations are examined based on the extended ordering in social choice theory. The equivalent conditions to the evaluations are deduced so that mathematical interpretations for the evaluations become possible, and order relations related to two evaluations are deduced.
This paper proposes the bootstrapped subjective evaluation model and the fuzzy measure identification method for its model. The proposed model is able to treat the interaction among the criteria empirically based on individual evaluation score. That can express a decision maker's preference than ordinary models such as additive model, Choquet integral model, cumulative prospect theoretical model and bi-capacity model. By comparing with the ordinary models, the numerical simulation and the discussion show the efficiency of the bootstrapped subjective evaluation model.
This paper proposes a method for quantitative evaluation of color harmony using a fuzzy reasoning. In the development of a system that evaluates color scheme quantitatively, it is important to clarify relationship between colors and their combination. In addition, it is also necessary to define some rules with relation to correlation between color scheme and color combination patterns. Therefore, this paper shows new approach to evaluate color harmony automatically. It uses some fuzzy sets to represent distribution of colors in hue and tone. Hue distribution types and tone distribution types are based on the color harmony theory proposed by Matsuda, and the system can evaluate harmony of three color combinations quantitatively using fuzzy rules based on the harmonious relationship between hue distribution type and tone distribution type. An experimental result shows that the result of quantitative evaluation using the system is similar to the questionnaire result about color harmony by human.
In this paper we discus influence of procedures for an interactive identification method on forms of identified membership functions. First, we clarify factors that influence forms of identified membership functions, taking the BASE method as an example. Secondly, we examine influence of three factors, 1) presentation order of elements, 2) number of categories for rating membership degree, 3) verbal labels of these categories, through two psychological experiments. The obtained results are as follows; 1) the presentation order affects transition area width of identified membership functions. 2) Especially, when the elements are presented in ascending order, the rating tends to become forced binary decision. 3) The verbal labels do not affect the forms of identified membership functions.
Identifying fuzzy measures from known partial data using the relative entropy is proposed. We consider an estimation with less relative entropy is better. We also propose the generalizations of λ-fuzzy measure and χ-fuzzy measure which are called (λ, p)-fuzzy measure and generalized χ-fuzzy measure, and show examples of applying our proposed identification to them.