• 한국컴퓨터정보학회:학술대회논문집
• /
• 한국컴퓨터정보학회 2023년도 제68차 하계학술대회논문집 31권2호
• /
• pp.47-50
• /
• 2023

본 논문에서는 딥러닝 CNN 알고리즘을 이용하여 사람의 얼굴 이미지를 학습한 다음 나이와 성별을 예측하는 시스템을 제안한다. 이 시스템은 개개인 마다 각기 다른 외형적 특성을 고려하여 이를 분석한 다음 이에 맞는 헤어 스타일, 옷차림을 추천할 수 있다. 해당 기술을 활용하여 메타버스 아바타 생성에 사용자의 얼굴과 같은 신체적 특성을 고려할 수 있다. 향후에는 신체 전체를 이미지화하여 보다 더 다양한 정보를 인식할 수 있도록 연구를 진행할 것이다.

• 한국콘텐츠학회논문지
• /
• 제20권8호
• /
• pp.10-18
• /
• 2020

본 연구에서는 콘텐츠 산업 중 음악 분야 2차 산업데이터를 활용하여 딥러닝 기법을 이용한 흥행 예측모델 구축 가능성을 살펴보았다. 본 연구를 통해 구축한 딥러닝 예측 모델은 17개 독립변인 -가수 파워, 가수 영향력, 피처링 가수 파워, 피처링 가수 영향력, 참여 가수 수, 참여 가수의 성별, 작사가 역량, 작곡가 역량, 편곡가 역량, 제작사 역량, 유통사 역량, 앨범의 타이틀 여부, 음원 스트리밍 플랫폼 좋아요 수, 음원 스트리밍 플랫폼 코멘트 수, 사전 홍보 기사 수, 티저 영상 조회 수, 초기 흥행성과를 기반으로 음원 흥행성과 -음원이 차트내 상주하는 기간을 예측하는 구조다. 추가적으로 본 연구가 딥러닝 기법을 콘텐츠 분야에 접목시킨 초기단계 연구임을 고려하여, 콘텐츠 흥행예측 선행연구에서 요인 추출을 위해 활용하는 선형회귀분석을 통해 변인 소거 후 구축한 DNN 예측모델과 예측률 비교를 진행하였다.

• 디지털융복합연구
• /
• 제19권9호
• /
• pp.257-269
• /
• 2021

논문은 심탄도(BCG, Ballistocardiogram) 센서를 이용하여 생체신호 데이터를 비침습, 무구속적인 방식으로 수집하고, ICT 기술과 고성능 컴퓨팅 환경에서 인공지능 기계학습 알고리즘을 활용하여 데이터 기반 혈당 예측 알고리즘 모델 개발 및 검증하는 방법을 제시하고 연구하는 것이다. 혈당수치 예측모델은 MLP 아키텍처에 입력노드는 심박수, 호흡수, 심박출량, 심박변이도, SDNN, RMSSD, PNN50, 나이, 성별이며, 은닉층 7개를 사용하였다. 실험 결과는 5회 실험한 학습데이터의 평균 MSE, MAE 및 RMSE 값은 각각 0.5226, 0.6328 및 0.7692이며 검증데이터 평균 값은 각각 0.5408, 0.6776, 0.7968이었으며, 결정계수(R2) 수치는 0.9997의 결과를 보였다. 데이터를 기반으로 한 혈당수치를 예측하는 모델을 표준화하고 데이터셋 수집과 예측 정확성을 검증하는 연구가 계속적으로 진행된다면 비침습 방식의 혈당 수준 관리에 활용될 수 있을 것으로 사료된다.

• 한국조사연구학회지:조사연구
• /
• 제8권1호
• /
• pp.55-79
• /
• 2007

본 연구에서는 2006년 5월에 KBS/SBS에서 공동으로 실시한 제4회 전국동시지방선거 출구조사 과정을 소개하는 동시에 출구조사 자료를 통해 투표자의 특성(성별 및 연령대)에 따른 투표성향과 시간대별 투표자 특성을 정리하고, 출구조사 선거구별 득표율을 예측치와 실제 개표결과를 비교하여 출구조사의 예측오차를 살펴본다. 이를 위해 이번 출구조사에서 적용된 전반적인 표본추출방법 및 실사과정을 정리하고, 출구조사에서 발생한 전체 오차를 편향과 표본추출오차로 분리하여 오차의 특성을 분석한다. 편향 발생원인 중 하나로 볼 수 있는 표본의 대표성을 검토하기 위해 출구조사 표본을 선관위에서 집계한 투표자 투표율 분석결과와 비교하여 평가한다. 아울러 이번 지방선거 출구조사에서 적용된 '층화 후 정렬계통추출법'에 따른 분산추정법을 적용하여 출구조사의 정확성을 표본추출오차 관점에서 살펴본다.

• 한국생활과학회지
• /
• 제13권5호
• /
• pp.659-672
• /
• 2004

This study examines different individual, family, and school environmental variables that affect victimization by peer aggression among adolescents. The sample consists of 868 seventh and eighth graders. Statistics and method for data analysis include Cronbach's alpha, percentage, means, standard deviation, Pearson correlation, multiple regression, and hierarchical regression. The major findings of this study are as follows: First, adolescents, both withdrawn and aggressive, have lower achievement in school work. Boys experience more direct victimization by peer aggression. Adolescents, especially boys, often experience indirect victimization by peer aggression, when they become withdrawn, own lower self-esteem, and have lower achievement in school work. Second, adolescents have more direct victimization by peer aggression when their parents are negligent of them. Also, adolescents seem exposed to indirect victimization by peer aggression when they receive more physical and emotional abuse and negligence from their parents. Third, adolescents experience more victimization by peer aggression-whether it's direct or indirect, when they cannot get adjusted to peer relations and get teachers' supervision. Fourth, as to direct victimization by peer aggression, withdrawal, one of the individual variables, is the most reliable prediction followed by gender, negligence, adaptability in peer relations, aggression, and teacher's supervision in sequence. For indirect victimization by peer aggression, withdrawal is the most reliable prediction followed by adaptability in peer relations, gender, physical and emotional abuse, and negligence in sequence.

• 한방비만학회지
• /
• 제17권2호
• /
• pp.68-76
• /
• 2017

Objectives: The purpose of this study was to analyze the factors affecting weight loss effect of Gamitaeumjowee-tang and to see if weight loss could be predicted using influence factors and weight loss progress. Methods: From September 2016 to March 2017, we retrospectively reviewed the medical records of 139 patients who were visited to the Korean Oriental Clinic for 3 months. We conducted a regression analysis to determine whether age, gender, initial weight, patient health questionnaire-9, heart rate variability (HRV), sleep quality, drinking habit and the medication history of weight loss affect weight loss. We found weight loss prediction equations using multiple regression analysis applying significant factors and weight loss progress. Results: Gender and initial weight had a significant effect on weight loss in all periods (P<0.001). HRV had a significant effect on primary weight loss (P<0.01). Other factors did not have any significant effect on weight loss. Using the significant factors, weight loss of each period could be predicted from 23.9% to 44.6%, and tertiary weight loss could be predicted with 76.6% using factors, primary weight loss and secondary weight loss (P<0.001). Conclusions: This study suggests that weight loss effect of Gamitaeumjowee-tang maybe be affected by influence factors and that weight loss prediction equations using them can be used for obesity treatment.

• 대한인간공학회지
• /
• 제31권5호
• /
• pp.695-704
• /
• 2012

Objective: The purpose of this study was to investigate major factors influencing adoption of smartphone to promote its use by older adults. Background: Despite increasing proportion of elderly people and elderly market, the proportion of elderly smartphone user is still relatively small compared to whole smartphone users. Thus, we need to find out major factors influencing adoption of smartphone to increase proportion of elderly smartphone users. Method: Seven major factors were extracted from 36 survey questions using factor analysis. Regression analysis was also applied to determine specific factors affecting intention of use based on user versus non-user of smartphone, age, gender, and educational background. Results: As results of factor analysis and regression analysis, major factors influencing adoption of smartphone for elderly users were significantly different according to gender, age, educational background based on smartphone users or non-users. Conclusion: The result of this study identified major factors influencing adoption of smartphone for the elderly and provided basic information related to adoption of smartphone according to elderly people's characteristics. Consequently, we can expect to reduce the information gap and to improve quality of life for the elderly. Application: The development and marketing strategy could be applied differently based on the factors influencing adoption of smartphone. It is also possible to develop a prediction model for smartphone adoption according to elderly users' characteristics.

• 보건교육건강증진학회지
• /
• 제14권1호
• /
• pp.97-113
• /
• 1997

It is very important to estimate the patients satisfaction level with medical services, to classify the objectvies according to the patients characteristics and sub-satisfaction factors. The purpose of this study is to determine the factors affecting satisfaction in oriental hospital. The 549 patients' hospitalized in five oriental hospital in Taegu city and one oriental hospital in Kyungbuk province were selected for this study. The results summarized are as follows. l. The general characteristics of 549 objectvies were included gender, age, education, occupation, income level, length of stay, health status of hospitalized, and expectation for medical care. 2. Patients characteristics affecting patients total satisfaction, as for age(b=0.05), health status of patients(b=-0.052), and expectation for medical care(b=0.117) were significant, while gender, education, job, income level, and length of stay were not. As the factors according to patients satisfaction, accessibility(b=0.09l), doctor's kindness(b=0.357), staff kindness(b=0.137), nurse's skills(b=0.111), hospital facilities(b=0.211), and medical fee(b=-0.160) were significant. In total patients' satisfaction, Doctor's kindness was the most significant of prediction variables. In general the factors affecting In-patient satisfaction of oriental hospital was highly associated with doctor's kindness.

• 보건교육건강증진학회지
• /
• 제27권3호
• /
• pp.75-84
• /
• 2010

Objectives: This study investigates factors influencing the perpetrations of sexual violence while dating among college students. Methods: With a correlational survey design, a self-report survey was conducted and collected 1,132 responses from college students with dating experiences. Methods including descriptive statistics, t-test, ANOVA, pearson correlation coefficients, and multiple regression were used to analyze data. Results: Compared to college students with no perpetrations of sexual violence, college students with perpetrations of sexual violence had significantly higher scores in father's violence, mother's violence, gender role stereotype, and sexual violence permissiveness. On the other hand they had significantly lower scores in sexual assault recognition than the compared group. A multiple regression model result forecasted parents' violence, sexual assault recognition, sexual violence permissiveness, and gender as prediction indicators of perpetrations of sexual violence. Conclusion: To prevent sexual violence while dating, domestic violence should be decreased through parents education and counseling from childhood. High-risk groups should be detected by surveying socio-psychological variables including experience of domestic violence, sexual assault recognition, and sexual violence permissiveness. It need to develop and implement sexual violence prevention programs to accurately inform and aware sexual violence.

• 대한가정학회지
• /
• 제45권6호
• /
• pp.11-20
• /
• 2007

The purpose of this study was to examine determine the relations relationship between the child's characteristics (i.e. temperament) and maternal behaviors (parenting and the content of demands) and the child's compliance. Data were gathered from 153 mothers of toddlers in Seoul, via questionnaires. The major principal results of this study were as follows: Neither gender nor age differences was found were observed in toddlers' compliance. Correlation analyses revealed significant relations relationships between both a child's emotionality and the mother's parenting, and a child's compliance. That is, in cases in which the child's emotional reactivity and the mother's authoritarian parenting were high, the child evidenced lower compliance, the child showed whereas the higher the mother's authoritative parenting were was, the higher compliance the child showed displayed. The hierarchical regressions analysis indicated that maternal demands on 'caring' was constituted the most significant variable to predict for the prediction of toddlers' compliance, and child activity level and maternal authoritative parenting behavior were also significant variables.