• Journal of Biomedical Engineering Research
• /
• v.20 no.3
• /
• pp.315-321
• /
• 1999

The purpose of this study is to develop a mathematical model for system identification in order to predIct muscle force based on eledromyographic signal. Therefore, a finding of the relalionship between characteristics of electromyographic signal and the corre spondng muscle force should be necessiiry through dynamic, joint model. To develop the dynamic joint model, the upper limb mcludmg the wrist and elbow joint has been considered. The kinematic and dynamic data, such as joint angular displacement, velocity, deceleration along with the moment of inertla, required to establish the dynamic model has been obtained by electrical flexible goniometer which has two degree-of-frcedoms. ln this model, muscle force can be predicted only electromyographs through the relationship between the integrated lorce and the mtegrated electromyographic signal over the duration of muscle contraclion in this study.

• Bulletin of the Korean Chemical Society
• /
• v.32 no.1
• /
• pp.251-262
• /
• 2011

Human ether-a-go-go related gene (hERG) potassium channel blockade, an undesirable side effect which might cause sudden cardiac death, is one of the major concerns facing the pharmaceutical industry. The purpose of this study is to develop an in silico QSAR model which uncovers the structural parameters of T-type calcium channel blockers to reduce hERG blockade. Comparative molecular similarity indices analysis (CoMSIA) was conducted on a series of piperazine and benzimidazole derivatives bearing methyl 5-(ethyl(methyl)amino)-2-isopropyl-2-phenylpentanoate moieties, which was synthesized by our group. Three different alignment methods were applied to obtain a reliable model: ligand based alignment, pharmacophore based alignment, and receptor guided alignment. The CoMSIA model with receptor guided alignment yielded the best results : $r^2$ = 0.955, $q^2$ = 0.781, $r^2_{pred}$ = 0.758. The generated CoMSIA contour maps using electrostatic, hydrophobic, H-bond donor, and acceptor fields explain well the structural requirements for hERG nonblockers and also correlate with the lipophilic potential map of the hERG channel pore.

• Journal of Korea Society of Industrial Information Systems
• /
• v.5 no.2
• /
• pp.47-55
• /
• 2000

To date many software engineering cost models have been developed to predict cost, schedule, and effort of the software under development. The COCOMO Ⅱ is well- suited for the new software development life cycle such as non-sequential and rapid- development processes. The traditional regression approach based on the least square criterion is the most commonly used technique for empirical calibration in the COCOMO Ⅱ model. It has a few assumptions frequently violated by software engineering data sets. The source data is also generally imprecise in reporting size effort, and cost-driver ratings, particularly across different organizations. And that the outlier for the source data is a peculiarity and indicates a data point. To cope with difficulties, in this paper, we propose a new regression method for calibrating COCOMO Ⅱ post-architecture model based on the minimum relative error(MRE) criterion. The characteristic of the proposed method is insensitive to the extreme values of the data in the empirical calibration. As the experimental results, It is evident that our proposed calibration method MRE was shown to be superior to the traditional regression approach for model calibration, as illustrated by the values obtained for standard deviation(^σ), and prediction at level LPRED(L) measures.

• The KIPS Transactions:PartD
• /
• v.12D no.5 s.101
• /
• pp.755-764
• /
• 2005

Recently, as it is considered more important to identify the function of ail unknown genes in living things, many tools for gene prediction have been developed to identify genes in the DNA sequences. Unfortunately, most of those tools use their own schemes to represent their programs results, requiring researchers to make additional efforts to understand the result generated by them So, it is desirable to provide a standardized method of representing predicted gene information, which makes it possible to automatically produce the predicted results for a given set of gene data In this paper, we describe an effective U representation for various predicted gene information, and present an XML-based analysis tool for gene predication results based on this representation. The developed system helps users of gene prediction tools to conveniently analyze the predicted results and to automatically produce the statistical results of the prediction. To show the usefulness of the tool, we applied our programs to the results generated by GenScan and GeneID, which are widely used gene prediction systems.

• Korean Journal of Agricultural Science
• /
• v.45 no.4
• /
• pp.823-837
• /
• 2018

The objective of this study was to evaluate the firmness and the sweetness index (SI) of tomatoes with a hyperspectral imaging (HSI) technique within the wavelength range of 1000 - 1550 nm. The hyperspectral images of 95 tomatoes were acquired with a push-broom hyperspectral reflectance imaging system, from which the mean spectra of each tomato were extracted from the regions of interest. The reference firmness and sweetness index of the same sample was measured and calibrated with their corresponding spectral data by partial least squares (PLS) regression with different preprocessing methods. The calibration model developed by PLS regression based on the Savitzky-Golay second-derivative preprocessed spectra resulted in a better performance for both the firmness and the SI of the tomatoes compared to models developed by other preprocessing methods. The correlation coefficients ($R_{pred}$) were 0.82, and 0.74 with a standard error of prediction of 0.86 N, and 0.63, respectively. Then, the feature wavelengths were identified using a model-based variable selection method, i.e., variable importance in projection, from the PLS regression analyses. Finally, chemical images were derived by applying the respective regression coefficients on the spectral image in a pixel-wise manner. The resulting chemical images provided detailed information on the firmness and the SI of the tomatoes. The results show that the proposed HSI technique has potential for rapid and non-destructive evaluation of firmness and the sweetness index of tomatoes.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.12
• /
• pp.3383-3397
• /
• 2023

Scene graphs serve as semantic abstractions of images and play a crucial role in enhancing visual comprehension and reasoning. However, the performance of Scene Graph Generation is often compromised when working with biased data in real-world situations. While many existing systems focus on a single stage of learning for both feature extraction and classification, some employ Class-Balancing strategies, such as Re-weighting, Data Resampling, and Transfer Learning from head to tail. In this paper, we propose a novel approach that decouples the feature extraction and classification phases of the scene graph generation process. For feature extraction, we leverage a transformer-based architecture and design an adaptive calibration function specifically for predicate classification. This function enables us to dynamically adjust the classification scores for each predicate category. Additionally, we introduce a Distribution Alignment technique that effectively balances the class distribution after the feature extraction phase reaches a stable state, thereby facilitating the retraining of the classification head. Importantly, our Distribution Alignment strategy is model-independent and does not require additional supervision, making it applicable to a wide range of SGG models. Using the scene graph diagnostic toolkit on Visual Genome and several popular models, we achieved significant improvements over the previous state-of-the-art methods with our model. Compared to the TDE model, our model improved mR@100 by 70.5% for PredCls, by 84.0% for SGCls, and by 97.6% for SGDet tasks.

• Tuberculosis and Respiratory Diseases
• /
• v.46 no.5
• /
• pp.654-661
• /
• 1999

Background : The upper respiratory tract is the primary target organ of various airborne pollutants and is easily accessible part of the respiratory tract, and also is the predominant structure where chronic cough originates. The nasal peak inspiratory flow(PIFn), which is the peak inspiratory flow via nose with nasal mask and spirometry, could be a reliable parameter of nasal obstruction. The validity of PIFn has been evaluated in several studies by assessing the correlation between PIFn measurements and other parameters of nasal air flow. This study was designed to show the reproducibility of PIFn, the difference of PIFn between patients with chronic cough and normal subjects, and the usefulness of PIFn in the evaluation of nasal obstruction in patients with chronic cough. Methods : PIFn was measured by spirometry with nasal mask, twice a day for 3 consecutive days in 7 young normal subjects to evaluate validity of the test. In 32 patients with chronic cough and 25 age-matched normal subjects, PIFn and pulmonary function test($FEV_1$, $FEV_1%$ pred, FVC, and FVC% pred) were measured at first visiting. Results : Values of PIFn, $FEV_1$, and FVC were nearly constant in 7 young normal adults. Patients with chronic cough were 32 (14 males and 18 females) and the mean age was $41.4{\pm}15.9$ years. Normal subjects were 32(22 males and 10 females) and the mean age was $39.8{\pm}18.6$ years. There was no significant difference of age and pulmonary function test between patients with chronic cough and normal subjects(p<0.05). The PIFn values in patients with chronic cough was significantly lower than those of normal subjects($2.25{\pm}0.68\;L/sec$ vs. $2.75{\pm}1.00\;L/sec$ ; p=0.02). The postnasal drip syndrome(PNDS) comprised the majority of patients with chronic cough(27). The PIFn in patients with PNDS was significantly lower than that of normal subjects (mean$\pm$SD ; $2.18{\pm}0.66$ vs. $2.75{\pm}1.00\;L/sec$, p=0.006). Conclusion : There was a significant difference of PIFn between patients with chronic cough and normal subjects. Among the patients with chronic cough, patients with PNDS showed the most significant difference with normal subjects in PIFn. The PIFn could be a useful parameter of nasal obstruction in patients with chronic cough, especially in patients with PNDS.

• Animal Bioscience
• /
• v.34 no.5
• /
• pp.867-879
• /
• 2021

Objective: Fibronectin 3 (FN3) and immunoglobulin like modules (Ig) are usually collocated beside modular cellulase catalytic domains. However, very few researches have investigated the role of these modules. In a previous study, we have sequenced and analyzed bacterial metagenomic DNA in Vietnamese goats' rumen and found that cellulase-producing bacteria and cellulase families were dominant. In this study, the properties of modular cellulases and the role of a FN3 in unique endoglucanase belonging to glycosyl hydorlase (GH) family 5 were determined. Methods: Based on Pfam analysis, the cellulases sequences containing FN3, Ig modules were extracted from 297 complete open reading frames (ORFs). The alkaline, thermostability, tertiary structure of deduced enzymes were predicted by AcalPred, TBI software, Phyre2 and Swiss models. Then, whole and truncated forms of a selected gene were expressed in Escherichia coli and purified by His-tag affinity column for assessment of FN3 ability to enhance enzyme activity, solubility and conformation. Results: From 297 complete ORFs coding for cellulases, 148 sequences containing FN3, Ig were identified. Mostly FN3 appeared in 90.9% beta-glucosidases belonging to glycosyl hydrolase family 3 (GH3) and situated downstream of catalytic domains. The Ig was found upstream of 100% endoglucanase GH9. Rarely FN3 was seen to be situated downstream of X domain and upstream of catalytic domain endoglucanase GH5. Whole enzyme (called XFN3GH5 based on modular structure) and truncate forms FN3, XFN3, FN3GH5, GH5 were cloned in pET22b (+) and pET22SUMO to be expressed in single and fusion forms with a small ubiquitin-related modifier partner (S). The FN3, SFN3 increased GH5 solubility in FN3GH5, SFN3GH5. The SFN3 partly served for GH5 conformation in SFN3GH5, increased modules interaction and enzyme-soluble substrate affinity to enhance SXFN3GH5, SFN3GH5 activities in mixtures. Both SFN3 and SXFN3 did not anchor enzyme on filter paper but exfoliate and separate cellulose chains on filter paper for enzyme hydrolysis. Conclusion: Based on these findings, the presence of FN3 module in certain cellulases was confirmed and it assisted for enzyme conformation and activity in both soluble and insoluble substrate.

• Journal of Science and Technology Studies
• /
• v.18 no.3
• /
• pp.153-216
• /
• 2018

The contemporary practices of Big-Data based automated decision making algorithms are widely deployed not just because we expect algorithmic decision making might distribute social resources in a more efficient way but also because we hope algorithms might make fairer decisions than the ones humans make with their prejudice, bias, and arbitrary judgment. However, there are increasingly more claims that algorithmic decision making does not do justice to those who are affected by the outcome. These unfair examples bring about new important questions such as how decision making was translated into processes and which factors should be considered to constitute to fair decision making. This paper attempts to delve into a bunch of research which addressed three areas of algorithmic application: criminal justice, law enforcement, and national security. By doing so, it will address some questions about whether artificial intelligence algorithm discriminates certain groups of humans and what are the criteria of a fair decision making process. Prior to the review, factors in each stage of data mining that could, either deliberately or unintentionally, lead to discriminatory results will be discussed. This paper will conclude with implications of this theoretical and practical analysis for the contemporary Korean society.

• Tuberculosis and Respiratory Diseases
• /
• v.42 no.3
• /
• pp.332-341
• /
• 1995

Background: Measurement of bronchodilator response is necessary to establish reversibility of airflow obstruction that was helpful to estimate the diagnosis, treatment, and prognosis in obstructive airway disease. An useful index should be able to detect the bronchodilator response more sensitively not related with degree of airflow obstruction and also be independent of initial $FEV_1$. Method: Sensitivities of bronchodilator response in each group classified by degree of airflow obstruction in $FEV_1$, FVC, $FEF_{25\sim75%}$, Isovolume $FEF_{25\sim75%}$, sGaw were studied and correlation coefficients were calculated between initial $FEV_1$ and reversibilities expressed as absolute, %initial, % predicted, %possible in $FEV_1$. Result: Sensitivities of bronchodilator response were 61.5% in FVC, Isovolume $FEF_{25\sim75%}$ and sGaw, in severe group, and 56.3% in Isovolume $FEF_{25\sim75%}$ and sGaw, in moderate group, and 62.5% in $FEV_1$ and sGaw and 50.0% in FVC and Isovolume $FEF_{25\sim75%}$, in mild group, and 60.0% in sGaw and 58.0% in Isovolume $FEF_{25\sim75%}$ in total patients. Correlation coefficients between initial $FEV_1$(L) and absolute, % initial, % predicted, % possible were 0.15, -0.22(p<0.05), 0.02, 0.24(p<0.05) and correlation coefficients between initial $FEV_1$(% predicted) and absolute, % initial, % predicted, %possible were 0.06, -0.28(p<0.05), 0.08, 0.39(p<0.05). Conclusion: Volume related parameters were more sensitive index not related with degree of airway obstruction and the change in $FEV_1$ expressed as % predicted was the least dependent on initial $FEV_1$ and reversibilities, expressed as % initial or as % possible(predicted minus initial $FEV_1$)were correlated with initial $FEV_1$.