• 한국컴퓨터정보학회논문지
• /
• 제25권7호
• /
• pp.39-46
• /
• 2020

최근, 대전 게임 분야에서, 가상 캐릭터들의 효과적인 설계를 위해 캐릭터의 승률을 효율적으로 예측할 수 있는 방법들이 매우 필요하다. 우리는 본 논문에서 이 문제를 해결하기 위해 시뮬레이션과 기계 학습을 결합하는 방법을 제안한다. 우선 대전 게임에서 가상 캐릭터의 대전 승률을 분석하기 위해서 시뮬레이션을 사용하고, 가상 캐릭터의 능력치에 따라서 승률을 예측하기 위해 회귀 모델에 기반한 기계 학습 기법을 적용한다. 제안한 기법으로 실험한 결과는 시뮬레이션 결과로 나온 승률과 기계 학습 기법이 예측한 승률이 거의 차이가 없다는 것을 확인하였다. 그리고 간단한 회귀 모델에 기반한 기계 학습으로도 실험에서 좋은 성능을 얻을 수 있었다.

• 품질경영학회지
• /
• 제34권2호
• /
• pp.129-135
• /
• 2006

In the advance of computer technology, it is possible to keep all the related informations for monitoring equipments in control and huge amount of real time manufacturing data in a data base. Thus, the statistical analysis of large data sets with hundreds of thousands observations and hundred of independent variables whose some of values are missing at many observations is needed even though it is a formidable computational task. A tree structured approach to classification is capable of screening important independent variables and their interactions. In a Six Sigma project handling large amount of manufacturing data, one of the goals is to screen vital few variables among trivial many variables. In this paper we have reviewed and summarized CART, C4.5 and CHAID algorithms and proposed a simple method of screening vital few variables by selecting common variables screened by all the three algorithms. Also how to develop a logistics regression model on a large data set is discussed and illustrated through a large finance data set collected by a credit bureau for th purpose of predicting the bankruptcy of the company.

• 한국농공학회논문집
• /
• 제55권6호
• /
• pp.47-55
• /
• 2013

The objective of this study is to evaluate the applicability of Communication, Ocean, and Meteorological Satellite (COMS) Geostationary Ocean Color Imager (GOCI) vegetation indices on a quantitative analysis. For evaluation, the vegetation indices such as RVI, NDVI and SAVI were extracted by using COMS GOCI and Terra Moderate Resolution Imaging Spectroradiometer (MODIS) imageries. The 4,000 points using simple random sampling (SRS) method were randomly extracted from land areas except ocean to compare the vegetation indices from two images. The results of linear regression showed that the regression coefficients of RVI, NDVI, and SAVI between COMS GOCI and Terra MODIS were 0.66~0.82, 0.71~0.83, and 0.71~0.83, respectively. Especially, the regression coefficients of RVI (r=0.85), NDVI (r=0.91) and SAVI (r=0.91) were strongly related from September 2011 to January 2012. Thus, COMS GOCI can be substituted for particular periods and it needs to verify additionally.

• Educational Technology International
• /
• 제15권2호
• /
• pp.71-88
• /
• 2014

A variety of studies to predict students' performance have been conducted since educational data such as web-log files traced from Learning Management System (LMS) are increasingly used to analyze students' learning behaviors. However, it is still challenging to predict students' learning achievement in blended learning environment where online and offline learning are combined. In higher education, diverse cases of blended learning can be formed from simple use of LMS for administrative purposes to full usages of functions in LMS for online distance learning class. As a result, a generalized model to predict students' academic success does not fulfill diverse cases of blended learning. This study compares two blended learning classes with each prediction model. The first blended class which involves online discussion-based learning revealed a linear regression model, which explained 70% of the variance in total score through six variables including total log-in time, log-in frequencies, log-in regularities, visits on boards, visits on repositories, and the number of postings. However, the second case, a lecture-based class providing regular basis online lecture notes in Moodle show weaker results from the same linear regression model mainly due to non-linearity of variables. To investigate the non-linear relations between online activities and total score, RF (Random Forest) was utilized. The results indicate that there are different set of important variables for the two distinctive types of blended learning cases. Results suggest that the prediction models and data-mining technique should be based on the considerations of diverse pedagogical characteristics of blended learning classes.

• 한국의류산업학회지
• /
• 제17권2호
• /
• pp.265-277
• /
• 2015

The aim of this study was to successfully reflect human body changes on the transformation of the virtual body within 3D virtual fitting spaces. For this purpose, existing problems of shape transformation of the virtual body were analyzed and regression equations which provides useful basic data for transformation of the virtual body that can be applied usefully to the 3D virtual fitting system was developed. Necessary data for the analyses were body measurement and 3D scan data of women with average physical form between the ages of 40 through 49. The reason that we used human body changes of the female subjects in their forties was based on the recognition that fundamental female body changes start to occur from age of forty. Body shapes were largely divided into 3 groups according to obesity which was found to be the biggest factor of shape change. Seven factors were extracted based on factor analysis of 47 body measurement categories and regression equations were created to extract specific measurements for each BMI group based on these seven factors. The major contribution of this paper can be summarized as follows. First, the regression equations to extract specific measurements based on the 7 representative variables remediated existing problem of virtual bodies as it increased the number of body shape transformation areas. Second, the regression equations helped to overcome the problem of current failing to reflecting changes in body cross-section shape based on simple girth measurements based on analysis of cross-section distances.

• 한국전산구조공학회:학술대회논문집
• /
• 한국전산구조공학회 2008년도 정기 학술대회
• /
• pp.184-189
• /
• 2008

In this study, an artificial wheel load generation method is proposed to assist practical engineers performing dynamic analysis with simple procedure. To generate an artificial wheel loads from running vehicle, PSD(Power Spectrum Density) profiles of actual wheel load were sampled in terms of various road roughnesses. A detailed truck and bridge models were used for sampling actual wheel load to represent the real motion of moving vehicle. These wheel load profiles were simplified for the artificial wheel load. The simplification of actual wheel load profiles was performed by regression analysis. The result showed that the artificial wheel load well represents the real profiles of wheel load.

• 한국의류학회지
• /
• 제25권5호
• /
• pp.913-924
• /
• 2001

The purpose of this study is to investigate the sensibility of body exposure in contemporary fashion. 31 kinds of costume samples have been selected from photographs in fashion magazines. Those were divided into part of exposed body, the way of exposure and kinds of exposure. I have measured them by using Semantic Differential method. The subjects were 120 male and 120 female in twenties. The obtained data were analyzed by factor analysis, ANOVA, MDS and regression analysis. The results of analysis are as follow; 1. Factor analysis has extracted four factors which consist of the sensibility of body exposure. These factors are Attraction, Activity, Boldness, Simplicity. 2. There were significant differences in design of body exposure and demographics. 3. Evaluative dimensions of the sensibility of body exposure were identified by Hard-Soft, Active-Elegant. 4. Preference was related to words which are buying desirable, attractive, simple, natural, active appealed by design of body exposure costume and buying desirable was related to words which are favorite, modest, beautiful, natural appealed by mood of exposure costume.

• International journal of advanced smart convergence
• /
• 제10권1호
• /
• pp.209-215
• /
• 2021

The purpose of this study is to determine the impact of broadcasters' broadcasting commentators' attributes on broadcasting conditions and the intention of continuous viewing of e-sports competitions. To achieve this goal, a survey was conducted on 300 university students who had experience watching e-sports competition broadcasting. The final 295 copies were selected as the final valid samples and data processing was carried out. Using SPSS 23, factor analysis, reliability analysis, correlation analysis, and simple and multiple regression analysis were performed. The result of this study was: First, reliability and attractiveness among the sub-factors of commentator attributes were shown to have a positive effect on broadcasting conditions. Second, reliability and attractiveness among the sub-factors of commentator attributes have a positive effect on the intent of continuous viewing. Third, broadcasting conditions have a positive effect on the intention of continuous viewing.

• 한국재난정보학회 논문집
• /
• 제15권2호
• /
• pp.249-258
• /
• 2019

연구목적: 본 연구에서는 서울 경기지역의 화강암을 대상으로 시추조사 시 채취된 암석 시료에서 측정된 물리적 특성 중 탄성파속도와 탄성계수로부터 암석의 일축압축강도와의 상관성을 분석하여 암석의 일축압축강도를 추정하는 것을 목적으로 한다. 연구방법: 119개의 화강암 시추 코아 시료를 대상으로 탄성파 속도와 탄성계수 그리고 일축압축강도와의 상관관계를 도출하기 위하여 실내 암석 실험을 실시하였다. 연구결과: 화강암에 대하여 탄성파속도와 일축압축강도 그리고 탄성계수와의 상호 관계를 단순회귀와 다중회귀로 분석한 결과는 전반적으로 관계식에 나타난 것처럼 신뢰도가 낮음을 확인하였다. 이는 화강암의 구성 입자가 균질하지 못하여 탄성파속도와 탄성계수를 이용한 압축강도 추정에 대한 활용성이 낮은 원인으로 사료된다. 결론: 본 연구에서 암석의 탄성파속도와 탄성계수로부터 일축압축강도를 추정하기 위한 상관관계를 분석하기 위하여 단순회귀분석과 다중회귀분석 방법을 이용하였다. 단순회귀분석은 결정계수($R^2$)가 0.61~0.67 이었으며, 다중회귀분석은 0.71로 나타났다. 따라서 다중회귀분석을 이용하여 암석의 일축압축강도를 추정할 때 다소 신뢰성이 높아질 수 있다. 또한, 향후 탄성파 속도와 탄성계수를 이용한 암석의 일축압축강도를 추정할 때 다양한 통계분석 기법(회귀분석, 인공신경망, 빅데이터 분석 등)을 활용하면 보다 신뢰성이 높은 결과를 얻을 수 있을 것으로 본다.

• 한국소음진동공학회논문집
• /
• 제26권7호
• /
• pp.806-813
• /
• 2016

To alleviate the vibration problem or to satisfy the required criteria for manifesting the guaranteed performance of precise equipment, various vibration isolation materials or apparatus, such as viscoelastic material, air and coil spring, have been developed and applied. Among them, a wire mesh material is regarded as one of the good candidate for reducing the vibration in terms of moderate material price, easy shape machining and long life cycle without the property deterioration induced by the aging or environmental effects. In this paper, prior to wire mesh isolator design, the static and dynamic elastic modulus of wire mesh materials are extracted from the experiment by the simple shaped cylindrical specimens and their characteristics for applying to vibration isolator design are examined. The simple shaped specimens were made as considering the design parameters of a wire mesh mount; i.e. the density, wire diameter and wire mesh slope, and the sensitivity analysis were also performed from a view point of the extracted elastic modulus.