• 환경영향평가
• /
• 제21권5호
• /
• pp.707-713
• /
• 2012

There are a lot of parameters affecting microbial acclimation/cultivation characteristics such as dynamic conditions, F/M ratio and substrate affinity. From the process control point of view, the effect of high salinity on the removal efficiencies of BOD and SS have been documented by few researchers. In this research, lab-scale CAS(Conventional Activated Sludge) process and modified $A_2O$(Anaerobic/Anoxic/Oxic) process were operated and monitored to evaluate the characteristics of microbial acclimation and cultivation under high salinity wastewater during the period of three weeks. As a result of acute microbial activity test(6hr) at various $Cl^-$ concentration, the appropriate $Cl^-$ concentration for microbial growth and acclimation ranged under 3,100 mg/l. As a result of acclimation/cultivation test, the trend of COD removal efficiency reduced gradually as time elapsed. It is considered that $NH_4$-N removal phenomenon of the conventional pollutants removal mechanisms gave little effect to the microbial acclimation/cultivation under high salinity wastewater.

• 대한의용생체공학회:학술대회논문집
• /
• 대한의용생체공학회 1995년도 추계학술대회
• /
• pp.173-176
• /
• 1995

We have developed a fast steady state free precession interferometry (SSFPI) technique which is useful for the fMRI (functional Magnetic Resonance Imaging). As is known, SSFP sequence with a suitable adjustment of gradient (readout) allows us to measure precession angle $\theta$ which is in turn related to the field inhomogeneity [1-3]. When the method is applied to the susceptibility effect based functional magnetic resonance imaging (fMRI), it was found that the direct susceptibility effect measurement was possible without perturbations such as the backgrounds and inflow effect. In this paper, simulation results and experimental results obtained with 2.0 Tesla MRI system are also presented.

• 한국컴퓨터정보학회:학술대회논문집
• /
• 한국컴퓨터정보학회 2023년도 제68차 하계학술대회논문집 31권2호
• /
• pp.529-532
• /
• 2023

본 논문은 소셜 네트워크 이용자의 텍스트 데이터를 대상으로, 트랜스포머 계열의 언어모델을 전이학습해 이용자의 MBTI 성격 유형을 분류한 국내 첫 연구이다. Kaggle MBTI Dataset을 대상으로 RoBERTa Distill, DeBERTa-V3 등의 사전 학습모델로 전이학습을 해, MBTI E/I, N/S, T/F, J/P 네 유형에 대한 분류의 평균 정확도는 87.9181, 평균 F-1 Score는 87.58를 도출했다. 해외 연구의 State-of-the-art보다 네 유형에 대한 F1-Score 표준편차를 50.1% 낮춰, 유형별 더 고른 분류 성과를 보였다. 또, Twitter, Reddit과 같은 글로벌 소셜 네트워크 서비스의 텍스트 데이터를 추가로 분류, 트랜스포머 기반의 MBTI 분류 방법론을 확장했다.

• Journal of Ginseng Research
• /
• 제37권4호
• /
• pp.457-467
• /
• 2013

A quick and simple method for simultaneous determination of the 30 ginsenosides (ginsenoside Ro, Rb1, Rb2, Rc, Rd, Re, Rf, Rg1, 20(S)-Rg2, 20(R)-Rg2, 20(S)-Rg3, 20(R)-Rg3, 20(S)-Rh1, 20(S)-Rh2, 20(R)-Rh2, F1, F2, F4, Ra1, Rg6, Rh4, Rk3, Rg5, Rk1, Rb3, Rk2, Rh3, compound Y, compound K, and notoginsenoside R1) in Panax ginseng preparations was developed and validated by an ultra performance liquid chromatography photo diode array detector. The separation of the 30 ginsenosides was efficiently undertaken on the Acquity BEH C-18 column with gradient elution with phosphoric acids. Especially the chromatogram of the ginsenoside Ro was dramatically enhanced by adding phosphoric acid. Under optimized conditions, the detection limits were 0.4 to 1.7 mg/L and the calibration curves of the peak areas for the 30 ginsenosides were linear over three orders of magnitude with a correlation coefficients greater than 0.999. The accuracy of the method was tested by a recovery measurement of the spiked samples which yielded good results of 89% to 118%. From these overall results, the proposed method may be helpful in the development and quality of P. ginseng preparations because of its wide range of applications due to the simultaneous analysis of many kinds of ginsenosides.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2021년도 추계학술발표대회
• /
• pp.626-628
• /
• 2021

사건 관계 추출 태스크는 구조화되지 않은 텍스트 데이터에서 사건의 구조화된 표현을 얻는 것이다. 하나의 문장에서도 많은 정보를 얻을 수 있는 중요한 태스크임에도 불구하고, 다양한 사전 학습 모델을 적용한 연구는 아직 활발하게 연구되지 않고 있다. 따라서 본 연구에서 사전 학습된 모델의 임베딩 기법 중 BERT, RoBERTa, SpanBERT에 각각 base, large 아키텍처를 적용하여 실험하였다. 사건을 식별하기 위한 trigger와 해당 trigger의 세부 argument를 식별하기 위한 분류기를 상위레이어로 각각 설계하였고, 다양한 배치 크기를 적용하여 실험하였다. 성능평가는 trigger/argument 각각 F1 score를 적용하였고, 결과는 RoBERTa large 모델에서 좋은 성능을 보인 것을 확인하였다.

• 대한방사선치료학회지
• /
• 제6권1호
• /
• pp.154-170
• /
• 1994

The Purpose of this study was to determine the degree and contents composing the qualify of life and to analyze the relationships among the demographic characteristics, the degree of pain and the quality of life of the cancer patients experiencing a radiation therapy. The subjects for this study were 110 out-patients experiencing a radation therapy at C University Hospital in K-city, from April to October, 1992. The data were obtained using a convenience sampling technique. The tool of this study was the quality of life scale developed by Ro, You-Ja and the data were analyzed using a SAS program for percentages, mean & standard deviation. ANOVA and Scheffe test. The results were as follows : 1. the average total score of the quality of life of the subjects was 139.65(minimum score 121-maximum score 164), item mean score grange 1-5) was 2.97. For each factor in the quality of life scale, the mean scores (range 1-5) were 3.29 in emotional state, 3.14 in relationship with neighbors. 3.04 in physical state and function. 2.92 in self-esteem, 2.81 in economic life and 2.65 in relationship with family. 2. The incidence of physical symptoms was seen fatigue ($84.5\%$). anorexia ($65.5\%$) and weight loss ($47.3\%$) in order. The 50.9 percentage of the subjects complained of moderate or severe pain. 3. The results of the analysis of the relationships between the demographic variables and the quality of life were as follows : Gender (F=8.45, P=0.0044), age (F=6.29, P=0.0001). educational level (F=5.67, P=0.0046), marital status (F=6.82. P=0.0016), occupation (F=2.86, P=0.009), monthly income (F=6.90, P=0.0003), family living together (F=3.95, P=0.0494) and person paying a medical fee (F=5.14, P=0.0023). 4. The relationship between the degree of pain and the quality of life was significant difference (F=3.12, P=0.0482).

• Archives of Pharmacal Research
• /
• 제28권2호
• /
• pp.190-194
• /
• 2005

In our search for monoamine oxidase (MAO) inhibitors from natural resources, we found that the methanol extract of the roots of Sophora flavescens showed an inhibitory effect on mouse brain monoamine oxidase (MAO). Bioactivity-guided isolation of the extract yielded two known flavonoids, formononetin (1) and kushenol F (2), as active compounds along with three inactive compounds, oxymatrine (3), trifolirhizin (4), and ${\beta}$-sitosterol (5). Formononetin (1) and kushenol F (2) showed significant inhibitory effects on MAO in a dose-dependent manner with $IC_{50}$ values of 13.2 and $69.9\;{\mu}M$, respectively. Formononetin (1) showed a slightly more potent inhibitory effect against MAO-B ($IC_{50}:\;11.0\;{\mu}M$) than MAO-A ($IC_{50}:\;21.2\;{\mu}M$). Kushenol F (2) also preferentially inhibited the MAO-B activity than MAO-A activity with the $IC_{50}$ values of 63.1 and $103.7\;{\mu}M$, respectively.

• 한국정보과학회 언어공학연구회:학술대회논문집(한글 및 한국어 정보처리)
• /
• 한국정보과학회언어공학연구회 2022년도 제34회 한글 및 한국어 정보처리 학술대회
• /
• pp.47-52
• /
• 2022

환자와 주변인들에게 다양한 문제를 야기하는 치매와 조현병 진단을 위한 모델을 제안한다. 치매와 조현병 진단을 위해 프로토콜에 따라 녹음한 의사와 내담자 음성 시료를 전사 작업하여 분류 태스크를 수행하였다. 사전 학습한 언어 모델의 MLM Head를 이용해 분류 태스크를 수행하는 Prompt 기반의 분류 모델을 제안하였다. 또한 많은 수의 데이터 수를 확보하기 어려운 의료 분야에 효율적인 Few-Shot 학습 방식을 이용하였다. CLS 토큰을 미세조정하는 일반적 학습 방식의 Baseline과 비교해 Full-Shot 실험에서 7개 태스크 중 1개 태스크에서 macro, micro-F1 점수 모두 향상되었고, 3개 태스크에서 하나의 F1 점수만 향샹된 것을 확인 하였다. 반면, Few-Shot 실험에서는 7개 태스크 중 2개 태스크에서 macro, micro-F1 점수가 모두 향상되었고, 2개 태스크에서 하나의 F1 점수만 향상되었다.

• 대한의용생체공학회:학술대회논문집
• /
• 대한의용생체공학회 1995년도 추계학술대회
• /
• pp.183-186
• /
• 1995

• Journal of the Optical Society of Korea
• /
• 제11권4호
• /
• pp.158-161
• /
• 2007

We proposed a novel linearization technique for a DFB LD in the RoF link. The proposed scheme is based on the cross gain modulation(XGM) effect of a reflective semiconductor optical amplifier(RSOA) with light injection. We experimentally demonstrated and evaluated the enhanced CIR performance using the proposed linearization scheme.