In the previous report, it was shown that for a standard hardness specimen, the relation between heat-treatment conditions and the indenting deformation process changes depending on the test conditions. In this report, the method of analysis was changed; aparameter design of the heat-treatment conditions and test conditions of the indenting deformation process was performed and the change caused by altering the two conditions was evaluated. This showed that the strength of the material is determined by the combination of the heat-treatment conditions and test conditions. By using the results of this indenting deformation process test, it became possible to study the meaning of the test as a strength of materials characteristic.
One of the purposes for which the MT system is used is identification. Usually identification is performed by judging from distances that indicate degrees of abnormality. As similarity can be considered as a degree of transferability of the true value and original information to the object, it was considered that transferability could be evaluated by using the concept of the S/N ratio of transferability. In the MT system, distances are obtained from matrix calculations that require more data than the number of items. This has been an obstacle in dealing with huge amounts of data, and attempts have been made to reduce the quantity of unit space data. To examine an application of the S/N ratio of transferability to the MT system and unit space data reduction, identification of seal impressions was chosen as an application of transferability evaluation and seal impression matching was evaluated by use of the standardized S/N ratio. The results show that seal impressions can be identified from the standardized S/N ratio value.
A method of using distances calculated by the Mahalanobis-Taguchi system (error root-mean-square method), based on the results of the classification of hazards according to GHS (Globally Harmonized System of Classification and Labelling of Chemicals), has been proposed as an index for showing the hazards posed by chemical substances in an integrated and intuitive way. In this report,the category data provided by the GHS hazard categories were numerized in several ways and an L18 orthogonal array was used to study the effects on the calculated distances of the setting of low hazard substances in the unit space, the treatment of unclassifiable substances (those with unknown properties,equivalent to missing data), the weighting of different hazards (selection of important hazards), and other factors. This led to a finding that in view of the quantity of substances in the unclassifiable category, a combination of selected substances and excluded substances was appropriate as the unit space, and that treating all unclassifiable substances as highly hazardous reduced the discriminative capabilities of the system. The tendencies of the numerization method and the weighting scheme to affect the calculated distances were also clarified.
Students taking a short course lecture in quality engineering were given problems without correct answers, five proposed criteria were used to grade their performance, and differences among the ratings were studied. The use of the criteria reduced variations in grades given by three instructors and enabled grading differences between the instructors to be reduced. In the calculation of the variation pressure of the grade evaluation, using a unit space constructed by the students with the average grade demonstrated a good agreement with the grading by the criteria. From the study of factorial effects of evaluation items on the variation pressure distance, it appears that grading variations among the instructors can be reduced by evaluating reports in which the students expressed their own ideas in the form of organized text according to the criteria. Evaluating students by using both the criteria and the variation pressure distance could become an advantageous method for many types of scholastic evaluation.