A SN ratio is famous as a criterion for parameter design using the Taguchi method. In recent years, it has been revealed that the statistical feature of the SN ratio is described by a doubly non-central F distribution. Accordingly, the characteristics of the doubly non-central F distribution are gaining attention again. However, it is extremely difficult to theoretically and exactly evaluate a percentile for the doubly non-central F distribution since its probability density function is very complicated. Therefore, approximation techniques for evaluating the percentile have been conventionally studied. One of them, a method for approximating the doubly non-central F distribution using central F distribution based on the first three approximate moments in the doubly non-central F distribution has been devised. Further, an evaluation of the doubly non-central F distribution percentile based on Cornish-Fisher expansion using those approximate values of first three moments has been proposed. In this study, we calculate the accurate values of moments in doubly non-central F distribution, and then propose a new method for evaluating the percentile of the distribution using the accurate values of the first two moments. Through some numerical simulations, the effectiveness of our new proposal in this study is confirmed. In addition, as an application of our proposal, the confidence interval of the SN ratio is evaluated.

View full abstract

In this paper, we focus on classification problems based on the vector space model. As one of the methods, distance metric learning which estimates an appropriate metric matrix for classification by using the iterative optimization procedure is known as an effective method. However, the distance metric learning for high dimensional data tends to cause the problems of overfitting to a training dataset and longer computational time. In addition, the number of parameters that need to be estimated is in proportion to the square of the input data dimension. Therefore, if the dimension of input data becomes high, the number of training data to acquire a metric matrix with enough accuracy becomes enormous. Especially, these problems are caused when analyzing the document data and purchase history data stored in the EC site with high dimensional and sparse structure. To avoid these problems, we propose the method of l₁ regularized distance metric learning by introducing the alternating direction method of multiplier (ADMM) algorithm. The effectiveness of our proposed method is clarified by classification experiments using a newspaper article that has a highly dimensional and sparse structure and the UCI machine learning repository, which has a low and dense structure.

View full abstract

In general, the quality of items produced in a manufacturing process is not necessarily uniform, and it has stochastic variability. Control charts are basic tools for detecting a change in process conditions using the quality data with stochastic variability. When the assignable change in process conditions is detected by the control charts, identifying the time point of state change and searching for assignable causes are important issues in process management. In this study, first we develop the state transition tracking method for depicting the state transition and showing the change points using successive observed data. Then, we discuss a new procedure for monitoring the process conditions using the state transtion tracking method with the control charts.

View full abstract

One of the most important problems in sports injuries. Recently, one issue is to reduce the incidence of sports injuries through advanced preventive measures regarding the leading risk factors of sports injuries. However, the criterion for these advanced preventive measures are not clear. The purposes of this paper are to determine the leading indexes of preference factors for preventing sports injuries in basketball and presenting these indexes to the basketball players or coaches using data with actual recorded basketball practice logs in order to construct a model for preventing sports injuries. In this paper, the onset and preventive factors for muscle and joint sports injuries in basketball are determined by logistic regression analysis and poisson regression analysis, which are typical analyses used for quantitative data. From the results of these analyses, we suggest preventive measures that decrease the incidence of sports injuries in basketball.

View full abstract

In recent years, it has been indicated that the statistical feature of the SN ratio, which is famous as a criterion for parameter design using the Taguchi method, is described by a doubly non-central F distribution. Accordingly, the characteristics of the doubly non-central F distribution are gaining attention again. It should be noted that the cumulative distribution function of doubly non-central F distribution is shown based on the incomplete beta function or Gauss hypergeometric function. Since the computations for incomplete beta function and Gauss hypergeometric function are complex, approximations for evaluating the cumulative distribution function of the doubly non-central F distribution have been investigated. In this study, a new approximation for the cumulative distribution function of the doubly non-central F distribution is proposed. Through some numerical validation, the features of the proposed approximation are addressed.

View full abstract

A training system for a visual inspection method utilizing peripheral vision is developed in this paper. It is said that the inspection method requires three factors: peripheral vision, instantaneous vision and saccadic eye movement. Specifically, in order to acquire the inspection method step by step, six training tools for three factors and their combinations are individually developed as computer system. Moreover, in order to verify the effectiveness of the training system, a control experiment with 12 subjects divided into two groups is implemented. As a result, it is observed that the training system improves the defect detection rate 20.4%.

View full abstract

This study aims to evaluate the relationship between physical workload and working posture during upward-pushing tasks in the supine position. For this purpose, we evaluated the pressure distribution and physical workload for each body part under various working conditions. For the experiment, subjects had to push each type of material with one hand under various conditions. There were nine total experimental conditions and the material was arranged for each condition. The horizontal distance of the material was set at three different levels (elbow, shoulder, and ear). The height of the material was set at three different levels (60%, 75%, and 90% of the length of the upper limb). We found that subjects assumed an upper-body position with the trunk inclined when the material was set higher. The pressure distribution indicated that the reaction force of the upper-body area was higher at the elbow than at the shoulder and ear levels. The reaction force of the head area was larger at heights of 60% and 90% of the length of the upper limb at elbow level. We found that the neck and trunk load increased when the height of the material was higher. The subjects assumed a posture with the upper-body inclined and the neck flexed forward less at both the level of the shoulder and the ear at a height of 75% of the length of the upper limb. The results of the scores indicated that the most comfortable condition was the height of 75% of the length of the upper limb. For the same conditions, the trunk and neck load were lower in this case than for others heights.

View full abstract

Recently, plastic bottle recycling firms in Japan are faced with the problem of instability in the procurement volumes of used plastic bottles during the bidding. On the other hand, there are many cases where used plastic bottles are collected using the reverse vending machine (RVM). Used plastic bottles that have been collected by RVMs are sent directly to recycling firms and recycled by the recycling firms. Therefore, RVMs provide more stable volumes when procuring used plastic bottles, and this process is more profitable for recycling firms as compared to the bidding process. We propose an evaluation model for introducing the RVM effect for plastic bottle recycling firms. At first, we construct a decision model for the bidding volume of the recycling firm, which incorporates three uncertainties the firm is facing. Then, we calculate the effect of introducing RVMs using the difference between values with/without RVMs. We show that there is a higher uncertainty for procurement volume when using the bidding process than there is if RVMs are introduced. As a result, we provide effective strategies for decision making regarding the procurement of RVMs.

View full abstract

In recent years, information systems are essential for decisions made by enterprises. However, there is one management challenge, as the cost of the maintenance process for the information system become larger than that of the system's development cost. One of the causes of this management challenge is the many inquiries that in the initial failure period after operating the information system. In previous studies, although factors of such causes have been studied, the cause remains unclear even after analysis. The purpose of this paper is to determine the factors that lead to inquiries in the initial failure by applying statistical analysis using actual collected data during the development of the information system. Moreover, a forecasting model for inquiries that occur in the initial failure period is proposed using factors for inquiries in the initial failure period.

View full abstract