• 응용통계연구
• /
• 제22권2호
• /
• pp.355-364
• /
• 2009

A simple method is proposed for constructing nonparametric confidence intervals for the difference or ratio of two median failure times. The method applies when clustered survival data with censoring is randomized either (I) under cluster randomization or (II) subunit randomization. This method is simple to calculate and is based on non-parametric density estimation. The proposed method is illustrated with the otology study data and HL-A antigen study data. Moreover, the simulation results are reported for practical sample sizes.

• 응용통계연구
• /
• 제23권6호
• /
• pp.1047-1056
• /
• 2010

임상시험에서 환자들을 각 처리로 할당할 때 반응변수에 영향을 미치는 공변량이 존재하면 공변량도 함께 고려하여 환자들을 랜덤하게 배치하여야 한다. 확률화(randomization) 방법들에는 여러 가지가 있으나 공변량에 따라 환자들을 배치하는 방법으로 층화(stratification)를 많이 사용한다. 층화는 환자들을 공변량에 따라 여러 층으로 나누고 각 층들 안에서 환자들을 랜덤하게 배치하는 방법인데, 공변량의 수가 많아지면 층의 수가 급격하게 늘어나기 때문에 층마다의 환자수가 충분히 많지 않으면 그 결과를 신뢰할 수 없게 된다. 이를 보완하기 위한 방법으로 Pocock과 Simon (1975)은 최소화(minimization) 방법을 제안하였으며 이 방법은 처리에 대한 공변량의 균형을 맞추는 것에 중점을 두었다. 본 논문에서는 현재 가장 많이 쓰이고 있는 확률화 방법들과 최소화 방법의 장단점, 불균형의 정도 및 검정력을 모의실험 연구를 통해 비교해보고자 한다.

• International Journal of Internet, Broadcasting and Communication
• /
• 제9권1호
• /
• pp.24-28
• /
• 2017

In the Internet of Things (IoT), resource-limited smart devices communicate with each other while performing sensing and computation tasks. Thus, these devices can be exposed to various attacks being launched and spread through network. For instance, attacker can reuse the codes of IoT devices for malicious activity executions. In the sense that attacker can craft malicious codes by skillfully reusing codes stored in IoT devices, code-reuse attacks are generally considered to be dangerous. Although a variety of schemes have been proposed to defend against code-reuse attacks, code randomization is regarded as a representative defense technique against code-reuse attacks. Indeed, many research have been done on code randomization technique, however, there are little work on analysis of the interactions between code randomization defenses and code-reuse attacks although it is imperative problem to be explored. To provide the better understanding of these interactions in IoT, we analyze how code randomization defends against code-reuse attacks in IoT and perform simulation on it. Both analysis and simulation results show that the more frequently code randomizations occur, the less frequently code-reuse attacks succeed.

• Genomics & Informatics
• /
• 제20권1호
• /
• pp.9.1-9.6
• /
• 2022

Mendelian randomization (MR) uses genetic variation as a natural experiment to investigate the causal effects of modifiable risk factors (exposures) on outcomes. Two-sample Mendelian randomization (2SMR) is widely used to measure causal effects between exposures and outcomes via genome-wide association studies. 2SMR can increase statistical power by utilizing summary statistics from large consortia such as the UK Biobank. However, the first-order term approximation of standard error is commonly used when applying 2SMR. This approximation can underestimate the variance of causal effects in MR, which can lead to an increased false-positive rate. An alternative is to use the second-order approximation of the standard error, which can considerably correct for the deviation of the first-order approximation. In this study, we simulated MR to show the degree to which the first-order approximation underestimates the variance. We show that depending on the specific situation, the first-order approximation can underestimate the variance almost by half when compared to the true variance, whereas the second-order approximation is robust and accurate.

• ETRI Journal
• /
• 제44권6호
• /
• pp.903-914
• /
• 2022

Mobile and telecommunication networking uses temporary and random identifiers (IDs) to protect user privacy. For greater intelligence and security o the communications between the core network and the mobile user, we design and build a dynamic randomization scheme for the temporary IDs for mobile networking, including 5G and 6G. Our work for ID randomization (ID-RZ) advances the existing state-of-the-art ID re-allocation approach in 5G in the following ways. First, ID-RZ for ID updates is based on computing, as opposed to incurring networking for the re-allocation-based updates, and is designed for lightweight and low-latency mobile systems. Second, ID-RZ changes IDs proactively (as opposed to updating based on explicit networking event triggers) and provides stronger security (by increasing the randomness and frequency of ID updates). We build on the standard cryptographic primitives for security (e.g., hash) and implement our dynamic randomization scheme in the 5G networking protocol to validate its design purposes, which include time efficiency (two to four orders of magnitude quicker than the re-allocation approach) and appropriateness for mobile applications.

• Journal of the Korean Statistical Society
• /
• 제25권1호
• /
• pp.39-48
• /
• 1996

A new method that uses information obtained not only from randomization device but also from direct question is introduced. The new maximum likelihood estimator is compared with those of Warner(1965) and Mangat(l994). For a choice of randomized device, we propose a choice depending on the sample size n and show that our estimator is more efficient than that of Mangat under the randomization device. The proposed procedure is extended to more general one which can be easily applied to some specific cases. Under the specified conditions, it is shown that the variance of this generalized estimator is smaller than that of Warner.

• Communications for Statistical Applications and Methods
• /
• 제21권1호
• /
• pp.81-91
• /
• 2014

This study derives the characteristic functions of (multivariate/generalized) t distributions without contour integration. We extended Hursts method (1995) to (multivariate/generalized) t distributions based on the principle of randomization and mixtures. The derivation methods are relatively straightforward and are appropriate for graduate level statistics theory courses.

• Journal of the Korean Statistical Society
• /
• 제14권1호
• /
• pp.48-55
• /
• 1985

In clinical trials subjects are avalible sequentially and must be assigned to treatments immediately. Completely randomized procedure for the allocation of treatments to each subject may result in severe imbalance among the number of subjects in treatment groups, especially for small experiments or interim analyses of large experiments. In this study, restricted randomization methods such as biased coin designs (Efron, 1971), permuted block design, and truncated binomial design are compared to teh completely randomized design in the presence of selection and/or accidential bias by Monte Carlo simulations.

• Communications for Statistical Applications and Methods
• /
• 제4권3호
• /
• pp.735-741
• /
• 1997

In the present paper, we propose a new randomization device free mail survey method. The estimator based on proposed model is unbiased and more efficient than the estimator based on SIngh, Mangat and Singh model (SMS-model)(1993) when $\pi$<1/2, and more protective than SMS-model in view of the protection of privacy regardless of the values of $\pi$ and $\pi_Y$ only if we count the number of say 'Yes' from the respondents. However, If we consider the respondents that say 'No', the SMS-model is more protective than our model.

• 대한예방의학회:학술대회논문집
• /
• 대한예방의학회 1998년도 제50차 추계 학술대회 연제집
• /
• pp.227-228
• /
• 1998