http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
머신러닝 기반 음성분석을 통한 체질량지수 분류 예측 - 한국 성인을 중심으로
김준호,박기현,김호석,이시우,김상혁,Kim, Junho,Park, Ki-Hyun,Kim, Ho-Seok,Lee, Siwoo,Kim, Sang-Hyuk 사상체질의학회 2021 사상체질의학회지 Vol.33 No.4
Objectives The purpose of this study was to check whether the classification of the individual's Body Mass Index (BMI) could be predicted by analyzing the voice data constructed at the Korean medicine data center (KDC) using machine learning. Methods In this study, we proposed a convolutional neural network (CNN)-based BMI classification model. The subjects of this study were Korean adults who had completed voice recording and BMI measurement in 2006-2015 among the data established at the Korean Medicine Data Center. Among them, 2,825 data were used for training to build the model, and 566 data were used to assess the performance of the model. As an input feature of CNN, Mel-frequency cepstral coefficient (MFCC) extracted from vowel utterances was used. A model was constructed to predict a total of four groups according to gender and BMI criteria: overweight male, normal male, overweight female, and normal female. Results & Conclusions Performance evaluation was conducted using F1-score and Accuracy. As a result of the prediction for four groups, The average accuracy was 0.6016, and the average F1-score was 0.5922. Although it showed good performance in gender discrimination, it is judged that performance improvement through follow-up studies is necessary for distinguishing BMI within gender. As research on deep learning is active, performance improvement is expected through future research.
김준호,Kim, Jun-Ho 한국종축개량협회 2006 種畜改良 Vol.11 No.4
본 글은 Veeppro dairy management 11월호의 내용을 발췌하여 정리한 내용입니다.
김준호,선순화,이석구,김정석 ( Joon Ho Kim,Soon Hwa Sun,Suk Koo Lee,Chung Suk Kim ) 한국산림과학회 1977 한국산림과학회지 Vol.35 No.1
The productive structure and the productivity of ×Populus albaglandulosa plantations, where are middle parts of the Korean peninsula, were studied by means of stratifying clip technique and of allometry. The densities of stands in the sample areas were 693 trees/ha in 6 year-old plantation and 527 or 625 trees/ha in 9 year-old one. The photosynthetic part of the productive structure was not shown normal conical form but layering. So this was efficient to transmit solar radiation into the stand floor. The standing crop of the terrestrial parts of 6 year-old plantation was 18.11 ton/ha and that of 9 year-old one 38.8 to 47.3 ton/ha. The wood volume to trunk to 6 year-old was 31.3㎥/ha and that of 9 year-old was 68.8 to 83.9㎥/ha. The annual net production of 6 year-old plantation was 4.8 ton/ha/year and that of 9 year-old one was 10.0 to 11.7 ton/ha/year and its wood volume of trunks was 17.9 to 21.1 ㎥/ha/year. In the 9 year-old plantation the standing crop or the annual net production was different between two sample areas. This seemed that the cause was not due to chemical character but to physical character of soil.
김준호,김동현,이상근,홍성경,정석영,Kim, Joon Ho,Kim, Dong Hyawn,Lee, Sang Geun,Hong, Seong Kyeong,Jeong, Sek Young 한국강구조학회 2009 韓國鋼構造學會 論文集 Vol.21 No.3
가스배관이 매설된 지역에서 지면고에 변화가 발생하는 공사 수행시 가스배관의 안정성을 확보하기 위한 위치이동을 수행한다. 본 논문에서는 위치이동에 따른 배관의 구조해석을 위한 모델링 방법의 최적화와 함께 위치이동의 단계별 발생 응력을 실시간으로 예측하기 위한 방법을 제안하였다. 모델링 방법으로는 요소의 종류와 크기, 배관 매설부의 경계조건 처리 방법, 세장비에 의한 기하학적 비선형 특성 등의 영향에 관하여 분석하였으며 정확성을 확보하면서 해석 효율을 높일 수 있는 조건을 구하였다. 배관의 응력 예측을 위해서는 위치이동의 수 단계에 발생하는 배관위치 및 최대응력 정보를 이용하여 인공신경망을 학습시켰으며 학습 후 세부 이동단계별 배관의 위치와 최대응력을 예측할 수 있도록 하였다. 개발된 응력예측시스템은 윈도우 환경의 프로그램으로 개발하였다. If there are some construction works that affect the stability of buried pipelines, the pipelines should be moved to guarantee their safety. In this paper, modeling methods for analyzing the movement of pipelines were sought, and the step-by-step stress estimation method of moving pipelines was developed. Some factors affecting of pipeline response such as the element type, the element size, boundary modeling, and geometric non-linearity were quantitatively investigated. In addition, some conditions in which accuracy and effectiveness can be compromised in the analysis of long pipelines were identified. A neural network was used to estimate the pipeline stress. The inputs to the neural network included step-by-step displacements, and the output was the resulting stress at each movement step. After training the neural network, it can be used to estimate pipeline stresses at some sub-steps that are not included in the training. A Windows-based stress estimation program was developed.