• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.1
• /
• pp.161-166
• /
• 2020

This paper suggests an extension of the unified Bayesian Tikhonov regularization method. The unified method establishes the relationship between Tikhonov regularization parameter and Bayesian hyper-parameters, and presents a formula for obtaining the regularization parameter using the maximum posterior probability and the evidence framework. When the dimension of the data matrix is m by n (m >= n), we derive that the total misfit has the range of m ± n instead of m. Thus the search range is extended from one to 2n + 1 integer points. Golden section search rather than linear one is applied to reduce the time. A new benchmark for optimizing relative error and new model selection criteria to target it are suggested. The experimental results show the effectiveness of the proposed method in the image restoration problem.

• Journal of Biomedical Engineering Research
• /
• v.22 no.1
• /
• pp.49-58
• /
• 2001

A new scale named LCL(Logically Classified Level) has been developed to judge the status of the spasticity quantitatively implementing a pcndulum test in this study. Total 30 parameters have been used to develop the new scale. One representative parameter that can represent the dominant characteristics of spasticity has been chosen through statistical analysis. 20 parameters among 30 parameters uscd in the statistical analysis were newly developed and 10 parameters were from previous studies. The new 20 parametcrs were developed using combinations of pcndulum test characteristics. ie anglc, angular velocity. musclo midel. and EMG. 11 parameters among 30 parameters have showed strong correlations each other, Finally. L11 that showed consistency at every case has been chosen to be a representative parameter among the 11 parameters. 28 patients data were separated into 4 groups. A regression equation to predict the trend of patients of L11 has been made. The paramcter L11 was tested to prove its usabilitics for various cases of patients. The new LCL scale is expected to be a quantitative scale, and to replace the MAS(Modified Ashworth Scale) that is not a quantitative scale. Especially it is also expected that the new scale could be used to plan a treatmcnt period. methods, and intensity. as it can evaluate the status of patient's in detail.

• Tuberculosis and Respiratory Diseases
• /
• v.58 no.3
• /
• pp.230-242
• /
• 2005

Background : Spirometry should be compared with the normal predictive values obtained from the same population using the same procedures, because different ethnicity and different procedures are known to influence the spirometry results. This study was performed to obtain the normal predictive values of the Forced Vital Capacity(FVC), Forced Expiratory Volume in 1 Second($FEV_1$), Forced Expiratory Volume in 6 Seconds($FEV_6$), and $FEV_1/FVC$ for a representative Korean population. Methods : Based on the 2000 Population Census of the National Statistical Office of Korea, stratified random sampling was carried out to obtain representative samples of the Korean population. This study was performed as a part of the National Health and Nutrition Survey of Korea in 2001. The lung function was measured using the standardized methods and protocols recommended by the American Thoracic Society. Among those 4,816 subjects who had performed spirometry performed, there was a total of 1,212 nonsmokers (206 males and 1,006 females) with no significant history of respiratory diseases and symptoms, with clear chest X-rays, and with no significant exposure to respiratory hazards subjects. Their residence and age distribution was representative of the whole nation. Mixed effect models were examined based on the Akaike's information criteria in statistical analysis, and those variables common to both genders were analyzed by regression analysis to obtain the final equations. Results : The variables affecting the normal predicted values of the FVC and $FEV_6$ for males and females were $age^2$, height, and weight. The variables affecting the normal predicted values of the $FEV_1$ for males and females were $age^2$, and height. The variables affecting the normal predicted values of the $FEV_1/FVC$ for male and female were age and height. Conclusion : The predicted values of the FVC and $FEV_1$ was higher in this study than in other Korean or foreign studies, even though the difference was < 10%. When compared with those predicted values for Caucasian populations, the study results were actually comparable or higher, which might be due to the stricter criteria of the normal population and the systemic quality controls applied to the whole study procedures together with the rapid physical growth of the younger generations in Korea.