• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.481-483
• /
• 2022

With the advent of the era of artificial intelligence, more and more fields have begun to use artificial intelligence technology, especially the medical field. Cancer is one of the biggest problems in the medical field. [1] If it can be detected early and treated early, the possibility of cure will be greatly increased. Malignant skin cancer, as one of the types of cancer with the highest fatality rate in recent years has problems such as relying on the experience of doctors and being unable to be detected and detected in time. Therefore, if artificial intelligence technology can be used to help doctors in early detection of skin cancer, or to allow everyone to detect skin lesions or spots anytime, anywhere, it will have great practical significance. In this paper we used attention residual learning convolutional neural network (ARL-CNN) model [2] to classify skin cancer pictures.

• Journal of Korean Society for Quality Management
• /
• v.36 no.3
• /
• pp.34-44
• /
• 2008

This research proposes a design method based on the statistical characteristics of the Shewhart control chart incorporated with 2 of 2 and 2 of 3 runs rules respectively. A Markov chain approach is employed in order to calculate the in-control and out-of-control average run lengths(ARL). Two different control limit coefficients for the Shewhart scheme and the runs rule scheme are derived simultaneously to minimize the out-of-control average run length subject to the reasonable in-control average run length. Numerical examples show that the statistical performance of the hybrid control scheme are superior to that of the original Shewhart control chart.

• Journal of Korean Society for Quality Management
• /
• v.33 no.3
• /
• pp.149-155
• /
• 2005

Cusum control chart is used for the purpose of controling the process mean. We consider the problem related to cusum chart for controling process variance. Previous researches have considered the same problem. The main difficulty shown in the related researches was to derive the ARL function which characterizes the properties of the chart. Sample variance, differently with sample mean, follows chi-squared type distribution, even when the quality characteristics are assumed to be normally distributed. The ARL function of cusum is described by a type of integral equation. Since the solution of the integral equation for non-normal distribution is not known well, people used simulation method instead of solving the integral equation directly, or approximation method by taking logarithm of the sample variance. Recently a new method to solve the integral equation for Erlang distribution was published. Here we consider the steps to apply the solution to the problem of controling process variance.

• Journal of Korean Society for Quality Management
• /
• v.32 no.4
• /
• pp.156-168
• /
• 2004

This research investigates the statistical efficiency of variable sampling size & sampling interval(VSSI) $\bar{X}$ charts under two assignable causes. Algorithms for calculating the average run length(ARL) and average time to signal(ATS) of the VSSI $\bar{X}$ chart are proposed by employing Markov chain method. States of the process are defined according to the process characteristics after the occurrence of an assignable cause. Transition probabilities are carefully derived from the state definition. Statistical properties of the proposed chart are also investigated. A simple procedure for designing the proposed chart is presented based on the properties. Extensive sensitivity analyses show that the VSSI $\bar{X}$ chart is superior to the VSS or VSI $\bar{X}$ chart as well as to the Shewhart $\bar{X}$ chart in statistical sense, even tinder two assignable causes.

• Journal of Korean Society for Quality Management
• /
• v.24 no.1
• /
• pp.73-86
• /
• 1996

Adaptive EWMA(Exponentially Weighted Moving Average)-x control chart using the Kalman gain recursive average is designed. The designed control chart is effective to on-line process monitoring as continuous flow processes. Performance evaluation between the designed control chart and traditional one is implemented. For this, ARL(Average Run Length) is adopted as a criterion. Results show that the designed adaptive EWMA-x control chart has shorter ARL than EWMA-x control chart when process mean is shifted. This model can be extended to process prevention control. The methodology proposed in this research is turned out to show the high performance than that of the given methodologies.

• Journal of Korean Society for Quality Management
• /
• v.22 no.1
• /
• pp.82-95
• /
• 1994

When independent individual measurements are taken both $S/c_4$ and $\bar{R}/d_2$ are unbiased estimators of the process standard deviation. However, with dependent data $\bar{R}/d_2$ is not an unbiased estimator of the process standard deviation. On the other hand $S/c_4$ is an asymptotic unbiased estimator. If there exists correlation in the data, positive(negative) correlation tends to increase(decrease) the ARL. The effect of using $\bar{R}/d_2$ is greater than $S/c_4$ if the assumption of independence is invalid. Supplementary runs rule shortens the ARL of X control charts dramatically in the presence of correlation in the data.

• Journal of the Korean Statistical Society
• /
• v.33 no.2
• /
• pp.219-231
• /
• 2004

Process adjustment is a complimentary tool to process monitoring in process control. Although original intention of process adjustment is not identifying a special cause, detection and elimination of special causes may lead to significant process improvement. In this paper, we examine the impact of special causes on process adjustment. The bias in the adjusted output process is derived for each type of special causes, and average run length (ARL) of the Shewhart chart applied to the adjusted output is computed for each special cause types. Numerical results are illustrated for the ARL of the Shewhart chart, thereupon seriousness of special causes on process adjustment is evaluated for each type of special causes.

• Industrial Engineering and Management Systems
• /
• v.16 no.1
• /
• pp.109-117
• /
• 2017

In this article, a control chart based on multiple dependent (or deferred) state sampling for the gamma distributed quality characteristic is proposed using the gamma to normal transformation. The proposed control chart has two pairs of control limits, which can be determined by considering the in-control average run length (ARL). The shift in the scale parameter of a gamma distribution is considered and the out-of-control ARL is evaluated. The performance of the proposed chart has been shown for different levels of the parameters of the proposed control chart. It is also shown that the proposed chart is better than the Shewhart chart in terms of ARLs. A case study with a real data has been included for the practical usage of the proposed scheme.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.30 no.4
• /
• pp.1-14
• /
• 2005

We propose a VSSI-CRL(Variable Sampling Size and Samplina Interval-Conforming Run length) synthetic control chart in order to improve the statistical characteristics of both the VSSI chart and the CRL synthetic chart. The VSSI-CRL chart utilizes VSSI sampling scheme, but it produces a signal only when the CRI is less than a given limit. An algorithm for calculating the ARL(Average Run length) and ATS(Average Time to Signal) of the VSSI-CRL chart is developed by employing Markov chain method. We present some lemmas for describing the statistical characteristics of the VSSI-CRL chart under in-control state. A procedure for designing the VSSI-CRL chart is proposed based on the lemmas. Extensive comparative studios show that the VSSI-CRL chart is superior to the CRL synthetic chart or the VSSI chart in general, and is comparable to the EWMA chart in ATS performance.

• Communications for Statistical Applications and Methods
• /
• v.5 no.3
• /
• pp.645-657
• /
• 1998

During the start-up of a process or in a job-shop environment conventional use of control charts may lead to erroneous results due to the limited number of subgroups used for the construction of control limits. This article considers the effect of using estimated control limits based on a limited number of subgroups. Especially we investigate the performance of $\overline{X}$ and R control charts when the data are independent, and X control chart when the data are serially correlated in terms of average run length(ARL) and standard deviation run length(SDRL) using simulation. It is found that the ARL and SDRL get larger as the number of subgroups used for the construction of the chart becomes smaller.