http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Influence on overfitting and reliability due to change in training data
Sung-Hyeock Kim,Sang-Jin Oh,Geun-Young Yoon,Yong-Gyu Jung,Min-Soo Kang 국제문화기술진흥원 2017 International Journal of Advanced Culture Technolo Vol.5 No.2
The range of problems that can be handled by the activation of big data and the development of hardware has been rapidly expanded and machine learning such as deep learning has become a very versatile technology. In this paper, mnist data set is used as experimental data, and the Cross Entropy function is used as a loss model for evaluating the efficiency of machine learning, and the value of the loss function in the steepest descent method is We applied the GradientDescentOptimize algorithm to minimize and updated weight and bias via backpropagation. In this way we analyze optimal reliability value corresponding to the number of exercises and optimal reliability value without overfitting. And comparing the overfitting time according to the number of data changes based on the number of training times, when the training frequency was 1110 times, we obtained the result of 92%, which is the optimal reliability value without overfitting.
박영수,서정훈,정재복,송시영,황영웅,장혁재,이준규,이진헌,이세준,정준표,강진경 대한소화기학회 2001 대한소화기학회지 Vol.38 No.4
Background/Aims: The only hope for the cure of pancreatic adenocarcinoma is curative resection. However, the rate of recurrence after curative resection is higher than 50%. The aim, of this study were to analyze survival and the factors influencing survival and to evaluate the recurrence rate and pattern after curative resection of pancreatic adenocarcinoma. Methods: The records of 250 patients who were diagnosed as pancreatic cancer were reviewed retrospectively. We classified the patients into 3 groups (curative resection, non-curative resection, and conservative treatment) and analyzed the factors influencing survival, recurrence rate, and recurrence pattern after curative resection. Results: Curative resection were performed in 31 (12.4%) of 250 cases. The patients with curative resection was significantly prolonged median survival time than those without non-curative resection or with conservative treatment. The factors influencing survival rate after curative resection were age, lymph node involvement, and disease status. Twenty-five of the 31 cases with curative resection had recurrences after 10 months. Local retroperitoneal recurrence was 69%. Liver metastasis and lymph node metastasis were occurred in 61% and 54%, respectively. Conclusions: The survival time of the patients with pancreatic adenocarcinoma was prolonged in the cases of curative resection compared with the cases of non-curative resection or conservative treatment. However, most cases after curative resection showed recurrence. Thus, early diagnosis and early treatment should be needed to prolong the survival time of the patients with pancreatic adenocarcinoma.
Influence on overfitting and reliability due to change in training data
Kim, Sung-Hyeock,Oh, Sang-Jin,Yoon, Geun-Young,Jung, Yong-Gyu,Kang, Min-Soo The International Promotion Agency of Culture Tech 2017 International Journal of Advanced Culture Technolo Vol.1 No.1
The range of problems that can be handled by the activation of big data and the development of hardware has been rapidly expanded and machine learning such as deep learning has become a very versatile technology. In this paper, mnist data set is used as experimental data, and the Cross Entropy function is used as a loss model for evaluating the efficiency of machine learning, and the value of the loss function in the steepest descent method is We applied the GradientDescentOptimize algorithm to minimize and updated weight and bias via backpropagation. In this way we analyze optimal reliability value corresponding to the number of exercises and optimal reliability value without overfitting. And comparing the overfitting time according to the number of data changes based on the number of training times, when the training frequency was 1110 times, we obtained the result of 92%, which is the optimal reliability value without overfitting.
인체 에너지 대사의 분석을 통한 효율적인 군 행군 속도에 관한 연구
김의수,임충희,서성혁,박두균 師範大學 體育硏究所 1991 서울大學校 體育硏究所論集 Vol.12 No.2
The purpose of this study is put to find a proper marching pace of Korean soldiers considering their general physical conditions. Subjects of 20 soldiers in service were sampled for this study, whose age, height and weight are 21.1 years, 171.6㎝, and 63.4kg, respectively. To achieve the purpose of this study, firstly, anaerobic threshold of the soldiers was measured to use as the maximum oxygen consumption. Secondly, oxygen consumption (Energy consumption) for different, marching paces was measured to compare with the maximum oxygen consumption measured under anaerobic threshold. And thirdly Oxygen consumption required for 1km marching distance at different paces measured and analyzed on an economic point of view. The outcome of the examination was follow: 1) Maximum Oxygen consumption of the soldiers was 3.83±5.7ml/kg/min.) and the anaerobic threshold was 2.73±0.22ℓ/min(43.4±5.2ml/kg/min) 2) Oxygen consumption for the speed of 4km/h was 46.2∼83.1% of anaerobic threshold. 3) Oxygen consumption for the speed of 4.5km/h was 54.1∼96.9% of anaerobic threshold. 4) Oxygen consumption for the speed of 5km/h was 67.9∼121.3% of anaerobic threshold. 5) Oxygen amount consumed for 1km marching distance was 18.81±2.38ℓ/km for the pace of 4km/h, 19.60±2.18/km for the pace of 4.5km/h, and 22.12±2.86ℓ/km for 5km/h. From the above illustrated results of the examination, 4.0∼4.5km/h was determined the most appropriate marching pace for Korean soldiers because it turned out to be the most economical pace without surpassing the maximum marching pace based on the Korean soldiers general physical conditions. Additionally, it is recommendable for commander in soldier not to have his soldiers sourpass the pace of 4.5km/h.
Deok Jea Cha,Dong Soo Kim,Sang Gil Lee,Si Hyeock Lee,Yong Bae Park 한국응용곤충학회 2016 한국응용곤충학회 학술대회논문집 Vol.2016 No.10
There are many other detection methods for Bursaphelenchus xylophilus. However, Baermann funnel method, PCR-based methods, which are laborious and time consuming processes that are unsuitable to rapid diagnose the Pine Wilt Disease (PWD) on the field. For these reasons, the aim of our experiment is not only to apply field diagnostic for pine wood nematode (PWN) but also to reduce total time for detection PWN in the pine trees by loop-mediated isothermal amplification (LAMP) with phosphate (Pi)-induced coloration reaction which could be yes or no answer for detection of PWN has not required UV detector but it just could be discriminated by naked eyes within 30 min. Genomic DNA was directly extracted from pine wood chips by Wood Chips Direct Lysis procedure which can be used for LAMP only after simple dilution. Our results suggest that LAMP-Pi detection method, simple and rapid method for detection of PWN, could be applied to the field diagnostic for PWD.
가정용치아미백제로서의 6% 과산화수소용액의 효과와 안전성
한수부,박상현,문혁수,Han, Soo-Boo,Park, Sang-Hyun,Moon, Hyeock-Soo 대한치주과학회 1994 Journal of Periodontal & Implant Science Vol.24 No.2
The purpose of this investigation was to study the efficacy and safety of 6% hydrogen peroxide gel as a daily home tooth bleaching gel. The subjects consisted of 20 male dental students representing a variety of acquired stain and each subject participated for a 4-week period. Tooth color analysis(Shade determination), sulcus bleeding index, probing depth and probing attachment level were done and recorded at baseline and at the end of each week of study. The results indicated that home bleaching gel containing 6% hydrogen peroxide was effective and caused no gingival inflammation. Sulcus bleeding index, probing depth and probing attachment level showed no change. In conclusion, 6% hydrogen peroxide gel is an effective and safe agent for daily home tooth bleaching.