• Proceedings of the KSPS conference
• /
• 2005.04a
• /
• pp.95-98
• /
• 2005

Many utterance verification (UV) algorithms have been studied to reject out-of-vocabulary (OOV) in speech recognition systems. Most of conventional confidence measures for UV algorithms are mainly based on log likelihood ratio test, but these measures take much time to evaluate the alternative hypothesis or anti-model likelihood. We propose a novel confidence measure which makes use of a momentary best scored state sequence during Viterbi search. Our approach is more efficient than conventional LRT-based algorithms because it does not need to build anti-model or to calculate the alternative hypothesis. The proposed confidence measure shows better performance in additive noise-corrupted speech as well as clean speech.

• Journal of the Korean Data and Information Science Society
• /
• v.9 no.1
• /
• pp.1-9
• /
• 1998

In case of Zero-Inflated Poisson model with a change-point, likelihood ratio test statistic was used for testing hypothesis for a change-point. A change-point and several interesting parameters were estimated by using the method of moments and maximum likelihood. In order to compare the estimators, empirical mean-square-error was used. Real data for the Zero-Inflated Poisson model with a change-point and Poisson model without a change-point were examined.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.23
• /
• pp.10525-10529
• /
• 2015

Objective: To investigate the relationship between the perception of cancer risk and likelihood of having undergone cancer screening. Materials and Methods: We used data from the Korean Health Panel Survey from December 2011 onward. Of 3,390 patients who visited a hospital during the previous year, we included data from 2,466 individuals; 924 samples were excluded due to missing data. Logistic regression analysis and the chi square test were used to investigate the association between perceived cancer risk and the likelihood of having undergone cancer screening. Results: For patients who perceived their risk of developing cancer during the next 10 years to be 30-40%, the odds ratio was increased 1.65 fold (95%CI: 1.223, 2.234) compared with those who perceived their risk to be almost zero. Although the difference was not statistically significant, perceiving cancer risk as either extremely low or extremely high appears to be associated with a reduced likelihood of having undergone cancer screening, resulting in an inverted U-shaped relationship. Conclusions: Physicians and researchers should be aware of the importance of the affective component of risk perception. Policies addressing the influence of cancer risk perception should be implemented in South Korea and worldwide.

• The Korean Journal of Applied Statistics
• /
• v.32 no.4
• /
• pp.561-572
• /
• 2019

We study false discovery rate segmentation (FDRSeg) and simultaneous multiscale change-point estimator (SMUCE) methods for multiscale multiple change-point estimation, and compare empirical behavior via simulation. FSRSeg is based on the control of a false discovery rate while SMUCE used for the multiscale local likelihood ratio tests. FDRSeg seems to work best if the number of change-points is large; however, FDRSeg and SMUCE methods can both provide similar estimation results when there are only a small number of change-points. As a real data application, multiple change-points estimation is done with the well-log data.

• Journal of Applied Reliability
• /
• v.7 no.2
• /
• pp.57-72
• /
• 2007

This paper suggests an approach for using ALTA 7 PRO to design accelerated life test plans. Conducting a accelerated life test requires finding life distributions at different stress levels and determining an appropriate life-stress relationship. Moreover, a test plan needs to be developed. In its optimal test plan, stress levels are determined and the proportions of test units are assigned at each stress level so that asymptotic variance of the maximum likelihood estimate of a (log) percentile of the life distribution at the design stress is minimized. Examples are presented for usage.

• Communications for Statistical Applications and Methods
• /
• v.11 no.3
• /
• pp.643-655
• /
• 2004

L$_1$-estimator in the linear regression model is widely recognized to have superior robustness in the presence of vertical outliers. While the L$_1$-estimation procedures and algorithms have been developed quite well, less progress has been made with the hypothesis test in the multiple L$_1$-regression. This article suggests computer-intensive resampling approaches, jackknife and bootstrap methods, to estimating the variance of L$_1$-estimator and the scale parameter that are required to compute the test statistics. Monte Carlo simulation studies are performed to measure the power of tests in small samples. The simulation results indicate that bootstrap estimation method is the most powerful one when it is employed to the likelihood ratio test.

• The Journal of the Acoustical Society of Korea
• /
• v.30 no.6
• /
• pp.324-329
• /
• 2011

In this paper, we propose a novel approach to improve the performance of a statistical model-based voice activity detection (VAD) which is based on the conditional maximum a posteriori (CMAP) with deviation. In our approach, the VAD decision rule is expressed as the geometric mean of likelihood ratios (LRs) based on adapted threshold according to the speech presence probability conditioned on both the speech activity decisions and spectral deviation in the pervious frame. Experimental results show that the proposed approach yields better results compared to the CMAP-based VAD using the LR test.

• Communications for Statistical Applications and Methods
• /
• v.22 no.6
• /
• pp.639-646
• /
• 2015

We propose simultaneous tests for mean and variance under the normality assumption. After formulating the null hypothesis and its alternative, we construct test statistics based on the individual p-values for the partial tests with combining functions and derive the null distributions for the combining functions. We then illustrate our procedure with industrial data and compare the efficiency among the combining functions with individual partial ones by obtaining empirical powers through a simulation study. A discussion then follows on the intersection-union test with a combining function and simultaneous confidence region as a simultaneous inference; in addition, we discuss weighted functions and applications to the statistical quality control. Finally we comment on nonparametric simultaneous tests.

• Journal of the Korean Data and Information Science Society
• /
• v.9 no.2
• /
• pp.95-103
• /
• 1998

Azzalini and Bowman proposed the pseudo-likelihood ratio test for checking the linear relationship using kernel regression estimator when the error of the regression model follows the normal distribution. We modify their method with the bootstrap technique to construct a new test, and examine the power of our test through simulation. Our method can be applied to the case where the distribution of the error is not normal.

• The Korean Journal of Applied Statistics
• /
• v.9 no.1
• /
• pp.53-73
• /
• 1996

This paper compares two accelerated life for Weibull distribution. One is the optimum constant stress accelerated life test which minimizes the asymptotic variance of maximum likelihood estimator of a specified quantile at design stress, and the other is corresponding simple step stress test. The models and optimum designs of constant stress and step stress tests are reviewed. Behaviors of asymptotic variances, effects of design parameters to optimum tests, and expected numbers of failures and expected test times of the two tests are investigated. The efficiency of step stress test relative to constant stress test is studied in terms of variance ratio, and robustness to preestimates of design parameters are investigated.