GACV for partially linear support vector regression

심주용,석경하 한국데이터정보과학회 2013 한국데이터정보과학회지 Vol.24 No.2

Partially linear regression is capable of providing more complete description of the linear and nonlinear relationships among random variables. In support vector regression (SVR) the hyper-parameters are known to affect the performance of regression. In this paper we propose an iterative reweighted least squares (IRWLS) procedure to solve the quadratic problem of partially linear support vector regression with a modified loss function, which enables us to use the generalized approximate cross validation function to select the hyper-parameters. Experimental results are then presented which illustrate the performance of the partially linear SVR using IRWLS procedure.

GACV for partially linear support vector regression

Shim, Jooyong,Seok, Kyungha The Korean Data and Information Science Society 2013 한국데이터정보과학회지 Vol.24 No.2

Partially linear regression is capable of providing more complete description of the linear and nonlinear relationships among random variables. In support vector regression (SVR) the hyper-parameters are known to affect the performance of regression. In this paper we propose an iterative reweighted least squares (IRWLS) procedure to solve the quadratic problem of partially linear support vector regression with a modified loss function, which enables us to use the generalized approximate cross validation function to select the hyper-parameters. Experimental results are then presented which illustrate the performance of the partially linear SVR using IRWLS procedure.

Modeling of subway indoor air quality using Gaussian process regression

Liu, Hongbin,Yang, Chong,Huang, Mingzhi,Wang, Dongsheng,Yoo, ChangKyoo Elsevier 2018 Journal of hazardous materials Vol.359 No.-

<P><B>Abstract</B></P> <P>Soft sensor modeling of indoor air quality (IAQ) in subway stations is essential for public health. Gaussian process regression (GPR), as an efficient nonlinear modeling method, can effectively interpret the complicated features of industrial data by using composite covariance functions derived from base kernels. In this work, an accurate GPR soft sensor with the sum of squared-exponential covariance function and periodic covariance function is proposed to capture the time varying and periodic characteristics in the subway IAQ data. The results demonstrate that the prediction performance of the proposed GPR model is superior to that of the traditional soft sensors consisting of partial least squares, back propagation artificial neural networks, and least squares support vector regression (LSSVR). More specifically, the values of root mean square error, mean absolute percentage error, and coefficient of determination are improved by 12.35%, 9.53%, and 40.05%, respectively, in comparison with LSSVR.</P> <P><B>Highlights</B></P> <P> <UL> <LI> Gaussian process regression (GPR) is proposed to predict particulate matter (PM<SUB>2.5</SUB>) in a subway. </LI> <LI> Compositional structure of different base kernels of GPR is proposed to model PM<SUB>2.5</SUB>. </LI> <LI> GPR with sum of squared-exponential and periodic kernels provides the best modeling performance. </LI> <LI> The proposed GPR was compared with other soft sensors including PLS, ANN, and SVM. </LI> </UL> </P>

양파의 생구무게 예측을 위한 여러 가지 일반회귀모형의 성능 비교

강윤정(Yunjeong Kang),나명환(Myung Hwan Na),조완현(Wanhyun Cho),고현석(Hyeon Seok Ko) 한국자료분석학회 2021 Journal of the Korean Data Analysis Society Vol.23 No.1

양파는 우리나라의 5대 채소 중 하나로 1인당 연간 소비량이 30kg에 달할 정도로 인기가 좋다. 양파의 수급은 전반적으로 국내 생산에 의존하므로 안정적인 가격으로 소비자에게 공급하기 위해서는 생산량 예측이 필요하다. 일반적으로 양파는 노지에서 재배되는데, 노지재배 특성상 급격한 기후 변화나 자연재해는 채소 생장에 큰 문제를 야기한다. 양파 재배 농가 및 관계기관에 서는 수확량 예측 및 증대를 위해 생육 최적의 환경 조건을 파악해야 하고 이를 위해 영향력이큰 환경 요인이 무엇인지 밝혀야 한다. 본 연구에서 우리는 주성분분석(PCA)과 부분최소제곱 (PLS)을 이용하여 재배기간 동안 양파의 생구 무게에 유의미한 영향을 주는 환경요소가 어떤 것인지 도출하고, 이를 통해 재배 시 생육 관리에 대한 전략을 제시하였다. 부분최소제곱 주성분이 현실적으로 우리가 예상하는 결과와 부합하였는데 생구 무게와 강수량, 습도는 음의 상관을 가지고 기온 및 지온의 차이, 일사량과는 양의 상관을 가짐을 확인하였다. 따라서 양파 생구의 안정적인 생장을 위해서는 습도, 일조, 기온 차에 관한 관리가 필요하다. 또한 일반 회귀모형인 주성분회귀모형(PCAR), 부분최소제곱회귀모형(PLSR) 그리고 서포트벡터머신회귀모형(SVMR)을 이용하여 수확 시기의 생구 무게를 예측하고 이들의 성능을 비교 해보았다. 서포트벡터머신회귀모 형이 월등히 우수한 성능을 보였고, 부분최소제곱회귀모형, 주성분 다항회귀 모형이었으나 큰 차이를 보이지는 않았다. Onions are one of Korea s five major vegetables and are so popular that the annual consumption per person reaches 30kg. Since the supply and demand of onions is entirely dependent on domestic production, it is necessary to predict the amount of production in order to supply them to consumers at a stable price. In general, onions are cultivated in the open field. Due to the characteristic of the field cultivation, rapid climate change or natural disasters cause a big problem in vegetable growth. Onion cultivation farms and related organizations must identify the optimal environmental conditions for growth in order to predict and increase the yield. In this study, we use principal component analysis (PCA) and partial least squares (PLS), to find environmental factors that have a significant influence on the weight of onions during the cultivation period, and through this, we present an environmental management strategy during cultivation. The partial least-squares principal component was realistically consistent with the results we expected. It was confirmed that the weight of bulbs, precipitation, and humidity had a negative correlation, and had a positive correlation with the difference between temperature and ground temperature, and the amount of insolation. Therefore, it is necessary to manage the difference in humidity, sunlight, and temperature for stable growth of onions. In addition, using the general regression model, the principal component regression model (PCAR), the partial least squares regression model (PLSR), and the support vector machine regression model (SVMR), we predicted the weight compared their performance. The model that predicted the weight of onion bulb best through experiments was the support vector machine, followed by the partial least squares regression model and the principal component polynomial regression model, but there was no significant difference.

Analysis of Carbonization Behavior of Hydrochar Produced by Hydrothermal Carbonization of Lignin and Development of a Prediction Model for Carbonization Degree Using Near-Infrared Spectroscopy(열수 탄화 공정을 거친 리그닌 하이드로차(hydrochar)의 탄화 거동 분석과 근적외선 분광법을 이용한 예측 모델 개발)

( Un Taek Hwang ),( Junsoo Bae ),( Taekyeong Lee ),( Sung-yun Hwang ),( Jong-chan Kim ),( Jinseok Park ),( In-gyu Choi ),( Hyo Won Kwak ),( Sung-wook Hwang ),( Hwanmyeong Yeo ) 한국목재공학회 2021 목재공학 Vol.49 No.3

In this paper, we investigated the carbonization characteristics of lignin hydrochar prepared by hydrothermal carbonization and established a model for predicting the carbonization degree using near-infrared spectroscopy and partial least squares regression. The carbon content of the hydrothermally carbonized lignin at the temperature of 200 ℃ was higher by approximately 3 wt% than that of the untreated sample, and the carbon content tended to gradually increase as the heating time increased. Hydrothermal carbonization made lignin more carbon-intensive and more homogeneous by eliminating the microparticles. The discriminant and predictive models using near-infrared spectroscopy and partial least squares regression approppriately determined whether hydrothermal carbonization has been applied and predicted the carbon content of hydrothermal carbonized lignin with high accuracy. In this study, we confirmed that we can quickly and nondestructively predict the carbonization characteristics of lignin hydrochar manufactured by hydrothermal carbonization using a partial least squares regression model combined with near-infrared spectroscopy. 본 논문에서는 열수 탄화(hydrothermal carbonization)에 의해 제조된 리그닌 하이드로차의 탄화 특성을 조사하였고, 근적외선 분광법과 부분 최소 제곱(partial least squares) 회귀를 이용하여 탄화 거동을 예측하기 위한 모델을 수립하였다. 온도 200℃에서 열수 탄화된 리그닌의 탄소 함량은 무처리 시료 보다 약 3 wt% 높았으며 가열 시간이 증가할수록 탄소 함량도 서서히 증가하는 경향이 나타났다. 열수 탄화는 리그닌을 더욱 탄소 집약적으로 변화시키고 마이크로 파티클을 제거하여 더욱 균질한 특성을 부여하였다. 근적외선 분광법과 부분 최소 제곱 회귀를 이용한 판별 및 예측 모델은 수열 탄화의 적용 여부를 완벽히 구분했으며 높은 정확도로 열수 탄화 리그닌의 탄소 함량을 예측하였다. 본 연구로부터 근적외선 분광법과 결합된 부분 최소 제곱 회귀 모델을 이용하여 열수 탄화에 의해 제조된 리그닌 하이드로차의 탄화 특성을 빠르고 비파괴적으로 예측할 수 있다는 것이 확인되었다.

스포츠 시장가치망 분류에 따른 파생시장이 일자리 창출에 미치는 영향 : 소셜데이터와 PLS(partial least squares regression) 회귀분석을 중심으로

한남희 한국사회체육학회 2020 한국사회체육학회지 Vol.0 No.81

Purpose: In this study, using the INSIGHT Deep MiningG program, frequency of mentioning and Emotion-Keyword are compared on a percentage basis to analyze the effect of a derived market of demand-oriented market value network classification on the creation of sports-related jobs. Method: Based on the preceding factors derived from social data analysis, the following conclusions were reached by using the method, partial least-square (PLS), to determine what influence derivative market has on the critical attributes of job creation. The study used the INSIGHT Deep MiningG program to analyze the impact of derived markets on the creation of sports-related jobs by using the demand-oriented market value chain classification as a percentage of social data references and emotional keywords. The following conclusions were reached using the PLS (PLS) regression method to determine the impact of job creation. Results: The results of this study are as follows. First, social users had the highest interest in the sports information market and the sports service market, given the amount of mention in the sports derivatives market of social data. Regarding sports job factors, interest in recruiting and job information was high. Second, the Emotion-Keyword of the sports derivative market showed high positive sentiment corresponding to the image of the new growth industry, while the Emotion-Keyword of the sports job showed negative sentiment due to the anxiety caused by issues such as unfair recruitments. Third, derivative market in the market value network mostly affected sports jobs. Conclusion: Job seekers prefer to work in the sports information market, sports service market, sports athlete training market, and sports tourism market, but the actual employment status is reflected in the sports information market, sports service market, sports tourism market, and sports facility market.

Milling tool wear forecast based on the partial least-squares regression analysis

Xu Chuangwen,Chen Hualing 국제구조공학회 2009 Structural Engineering and Mechanics, An Int'l Jou Vol.31 No.1

Power signals resulting from spindle and feed motor, present a rich content of physical information, the appropriate analysis of which can lead to the clear identification of the nature of the tool wear. The partial least-squares regression (PLSR) method has been established as the tool wear analysis method for this purpose. Firstly, the results of the application of widely used techniques are given and their limitations of prior methods are delineated. Secondly, the application of PLSR is proposed. The singular value theory is used to noise reduction. According to grey relational degree analysis, sample variable is filtered as part sample variable and all sample variables as independent variables for modelling,and the tool wear is taken as dependent variable, thus PLSR model is built up through adapting to several experimental data of tool wear in different milling process. Finally, the prediction value of tool wear is compare with actual value, in order to test whether the model of the tool wear can adopt to new measuring data on the independent variable. In the new different cutting process, milling tool wear was predicted by the methods of PLSR and MLR (Multivariate Linear Regression) as well as BPNN (BP Neural Network) at the same time. Experimental results show that the methods can meet the needs of the engineering and PLSR is more suitable for monitoring tool wear.

벌점함수와 부분최소자승법을 이용한 분류 방법

김윤대(Kim, Yundae),이혜선(Lee, Hyeseon),전치혁(Jun, Chi-Hyuck) 대한산업공학회 2011 대한산업공학회 춘계학술대회논문집 Vol.2011 No.5

Classification is to generate a rule of classifying objects into several categories based on the learning sample. Good classification model should classify new objects with low misclassification error. Many types of classification methods have been developed including logistic regression, discriminant analysis and tree. This paper presents a new classification method using penalized partial least squares. Penalized partial least squares can make the model more robust and remedy multicollinearity problem. This paper compares the proposed method with logistic regression and partial least squares classification methods by some real data. It is concluded that the new method has better power as compared with other methods.

벌점함수와 부분최소자승법을 이용한 분류 방법

김윤대(Kim, Yundae),이혜선(Lee, Hyeseon),전치혁(Jun, Chi-Hyuck) 한국경영과학회 2011 한국경영과학회 학술대회논문집 Vol.2011 No.5

Classification is to generate a rule of classifying objects into several categories based on the learning sample. Good classification model should classify new objects with low misclassification error. Many types of classification methods have been developed including logistic regression, discriminant analysis and tree. This paper presents a new classification method using penalized partial least squares. Penalized partial least squares can make the model more robust and remedy multicollinearity problem. This paper compares the proposed method with logistic regression and partial least squares classification methods by some real data. It is concluded that the new method has better power as compared with other methods.

A Simulation Study on the Prediction Ability of Biased Regression Methods

Jong-Duk Kim 한국자료분석학회 2002 Journal of the Korean Data Analysis Society Vol.4 No.4

This article examines three biased methods used for predictive modelling: ridge regression, principal component regression and partial least squares regression. A Monte Carlo study is performed to compare the prediction ability of these methods along with ordinary least squares. 48 different situations are considered and the root mean squared error of prediction is computed by using leave-one-out cross-validation for the comparison. Performance of the four methods are compared in detail.