The evaluation result of the commodities in daily life has an important meaning for Kansei engineering. However, neither the particular evaluation technique nor the analysis technique has been so researched. The “graphical modeling”is applied to analyze the time series judgments in the Kansei and sensory evaluation. This is a technique to which the pass analysis is done with the repeated calculation. Three kinds of same maker's chewing gums are used as stimuli. They are muscat taste, Japanese apricot taste, and bubble gum. The five items in each stimulus that are evaluated with a 5-point rating scale are used, for example, sweetness, fl avor, palatability, and so on. The causal relations of all 15 kinds are analyzed with the results of ten subjects. In the time series of evaluation values, the time order always corresponds to the causal relation. Only the pass coefficient from the evaluation of the previous state to the following evaluation is calculated. A comparatively strong causal relation is admitted to the evaluation item that relates to the characteristic of each stimulus. It is suggested that the Kansei and sensory evaluation of an important quality of the commodities is not to become independent reciprocally in the time series. This means that the previous evaluation exerts the influence until the future

View full abstract

A subjective weight sensation was able to express as a physical information quantity with the aid of the well-known Shannon's information theory in order to estimate Kansei information processing in weight perception of colored balls. The Kansei information for the weight sensation of colored balls is obtained from combinations of actions such as looking, hefting, hearing and so on. In the weight sensation, the contribution of the Kansei information got from each action was quantitatively analyzed by calculating the Shannon's information quantity.
The experiment found that for all methods except H judgment subjects felt that colors with low reflectivity such as black and purple were heavier, whereas colors with high reflectivity such as white and yellow were lighter. It was also found that the absolute value of the perception of weight tended to be smaller when the subjects made their judgments as they held the balls.
The differences among the different weight judgment conditions were expressed quantitatively by the amount of information communicated to perception. In I judgment, the information on the sensation of weight communicated was 0.60 bit, whereas it was reduced to 0.16 bit under I&H judgment. Similarly, the information communicated under L judgment was 0.43 bit, but was reduced to 0.12 bit under L&H judgment.
The above findings suggested the possibility of quantitatively presenting any differences according to the number and combination of different senses involved in a judgment of weight, and controlling the perception of the weight of an object in virtual space through color and luster.

View full abstract

One of the most important roles of Kansei engineering is to understand and operate the human tacit knowledge. The neural network approach is very useful for the learning of tacit knowledge. The neural network approach makes the tacit knowledge operational, and we can use tacit knowledge effectively. Since the learned knowledge is embedded in the neural networks, the l earned knowledge is opaque and difficult to understand. And the leaned knowledge itself is difficult to evaluate. Hence the generation of if-then rules from neural networks is proposed, the generated rules are usually related to many cells, and the rules are not easy to understand. Especially, there are many hidden cells in BP algorithm, and the meanings of the hidden cells are difficult to identify. If we want to understand the learned knowledge from neural networks, we need some way of interpretation of the learned k nowledge to the formal knowledge. In this report, we apply revised learning algorithms from BP algorithm. Those algorithms are useful to clarify the relationships among the cells and the behaviors of hidden cells. And then we present a methodology of interpretation of the knowledge learned by the neural networks to the formal knowledge that can be easy to understand. We discuss the steps of knowledge interpretation from neural network, and show the Kansei engineering example of the knowledge interpretation, and discuss the role and the possibility of the knowledge interpretation.

View full abstract