• International Journal of Reliability and Applications
• /
• v.11 no.1
• /
• pp.17-22
• /
• 2010

It is desired to estimate the reliability of a system that has two subsystems connected in series where each subsystem has two components connected in parallel. A batch sequential sampling scheme is introduced. It is shown that the batch sequential sampling scheme is asymptotically optimal as the total number of units goes to infinity. Numerical comparisons indicate that the batch sequential sampling scheme performs better than the balanced sampling scheme and is nearly optimal.

• Communications for Statistical Applications and Methods
• /
• v.12 no.1
• /
• pp.181-189
• /
• 2005

This paper suggest a method of inclusion probability proportional to size sampling that provides a non-negative and stable variance estimator. The sampling procedure is quite simple and flexible since a sampling design is easily obtained using mathematical programming. This scheme appears to be preferable to Nigam, Kumar and Gupta's (1984) method which uses a balanced incomplete block designs. A comparison is made with their method through an example in the literature.

• Survey Research
• /
• v.12 no.2
• /
• pp.77-88
• /
• 2011

Korean telephone surveys have been based on telephone directory and thus criticized for considerable under-coverage. Now, Korean survey institutions progress to random digit dialing (RDD) very actively. But still most surveys are administered by quota sampling, prone to assign heavier weights to social classes with more hours staying indoor. As a practical remedy, time-balanced quota sampling scheme was proposed by Huh and Hwang (2006). This study compares two telephone surveys on TV audience environment in Korea: RDD with conventional quota sampling versus RDD with time-balanced quota sampling.

• Proceedings of the Korean Statistical Society Conference
• /
• 2003.10a
• /
• pp.123-130
• /
• 2003

We define a broad class of rotation designs whose monthly sample is balanced in interview time, level of recall, and rotation group, and whose rotation scheme is time-invariant. The necessary and sufficient conditions are obtained for such designs. Using these conditions, we derive a minimum variance unbiased generalized composite estimator (MVUGCE). To examine the existence of time-in-sample bias and recall bias, we also propose unbiased estimators and their variances. Numerical examples investigate the impacts of design gap, non-sampling error sources, and two types of correlations on the variance of MVUGCE.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.7
• /
• pp.311-316
• /
• 2019

Classification problem is to predict the class to which an input data belongs. One of the most popular methods to do this is training a machine learning algorithm using the given dataset. In this case, the dataset should have a well-balanced class distribution for the best performance. However, when the dataset has an imbalanced class distribution, its classification performance could be very poor. To overcome this problem, we propose an over-sampling scheme that balances the number of data by using Conditional Generative Adversarial Networks (CGAN). CGAN is a generative model developed from Generative Adversarial Networks (GAN), which can learn data characteristics and generate data that is similar to real data. Therefore, CGAN can generate data of a class which has a small number of data so that the problem induced by imbalanced class distribution can be mitigated, and classification performance can be improved. Experiments using actual collected data show that the over-sampling technique using CGAN is effective and that it is superior to existing over-sampling techniques.