• Asia pacific journal of information systems
• /
• 제11권1호
• /
• pp.61-73
• /
• 2001

Predicting a variable using other variables in a large data set is a very difficult task. It involves selecting variables to include in a model and determining the shape of the relationship between variables. Nonparametric regression such as smoothing splines and neural networks are widely-used methods for such a task. We propose an alternative method based on a genetic algorithm(GA) to solve this problem. We applied GA to regression splines, a nonparametric regression method, to estimate functional forms between variables. Using several simulated and real data, our technique is shown to outperform traditional nonparametric methods such as smoothing splines and neural networks.

• Communications for Statistical Applications and Methods
• /
• 제31권3호
• /
• pp.323-336
• /
• 2024

The accurate forecasting of insurance claims is a critical component for insurers' risk management decisions. Hierarchical Bayesian parametric (BP) models can be used for health insurance claims forecasting, but they are unsatisfactory to describe the claims distribution. Therefore, Bayesian nonparametric (BNP) models can be a more suitable alternative to deal with the complex characteristics of the health insurance claims distribution, including heavy tails, skewness, and multimodality. In this study, we apply both a BP model and a BNP model to predict group health claims using simulated and real-world data for a private life insurer in Indonesia. The findings show that the BNP model outperforms the BP model in terms of claims prediction accuracy. Furthermore, our analysis highlights the flexibility and robustness of BNP models in handling diverse data structures in health insurance claims.

• Communications for Statistical Applications and Methods
• /
• 제2권2호
• /
• pp.266-276
• /
• 1995

We have considered the study of local influence for smoothing parameter estimates in nonparametric regression model. Practically, generalized cross validation(GCV) does not work well in the presence of data perturbation. Thus we have proposed local influence measures for GCV estimates and examined effects of diagnostic by above measures.

• 대한안전경영과학회지
• /
• 제9권3호
• /
• pp.165-171
• /
• 2007

Dependent models in quality statistics are classified as serially autocorrelated model, multivariate model and dependent sample model. Dependent sample model is most efficient in time and cost to obtain samples among the above models. This paper proposes to implement parametric and nonparametric models into production system depended on demand pattern. Nonparametric models have distribution free and asymptotic distribution free techniques. Quality statistical models are classified into two categories ; the number of dependent sample and the type of data. The type of data consists of nominal, ordinal, interval and ratio data. The number of dependent sample divides into 2 samples and more than 3 samples.

• Communications for Statistical Applications and Methods
• /
• 제22권5호
• /
• pp.463-473
• /
• 2015

We consider a nonparametric AR(1) model with nonparametric ARCH(1) errors. In order to estimate the unknown function of the ARCH part, we apply the stationary bootstrap procedure, which is characterized by geometrically distributed random length of bootstrap blocks and has the advantage of capturing the dependence structure of the original data. The proposed method is composed of four steps: the first step estimates the AR part by a typical kernel smoothing to calculate AR residuals, the second step estimates the ARCH part via the Nadaraya-Watson kernel from the AR residuals to compute ARCH residuals, the third step applies the stationary bootstrap procedure to the ARCH residuals, and the fourth step defines the stationary bootstrapped Nadaraya-Watson estimator for the ARCH function with the stationary bootstrapped residuals. We prove the asymptotic validity of the stationary bootstrap estimator for the unknown ARCH function by showing the same limiting distribution as the Nadaraya-Watson estimator in the second step.

• 문화기술의 융합
• /
• 제8권1호
• /
• pp.531-536
• /
• 2022

본 논문은 꼬리가 두꺼운 분포의 꼬리부분에 대한 분포를 추정할 경우 모수적 방법과 비모수적 방법의 성능에 대해 비교하였다. 모수적 방법으로는 일반화 극단값 분포와 일반화 파레토 분포를 이용하였고, 비모수적 방법은 커널형 확률밀도함수 추정방법을 적용하였다. 두 접근법의 비교를 위해 2014년부터 2018년까지 서울시 관측소별 일일 미세먼지 공공데이터를 이용하여 블록 최댓값 모형과 분계점 초과치 모형을 적용하여 함수 추정한 결과를 함께 보이고 2년, 5년, 10년의 재현수준을 통해 고농도의 미세먼지가 일어날 지역을 예측하였다.

• Journal of the Korean Statistical Society
• /
• 제30권4호
• /
• pp.645-660
• /
• 2001

In this paper we propose a goodness-of-fit test statistic for testing the (null) parametric model versus the (alternative) nonparametric model. Most of existing nonparametric test statistics are based on the residuals which are obtained by regressing the data to a parametric model. Our test is based on the bootstrap estimator of the probability that the smoothing parameter estimator is infinite when fitting residuals to cubic smoothing spline. Power performance of this test is investigated and is compared with many other tests. Illustrative examples based on real data sets are given.

• Journal of the Korean Statistical Society
• /
• 제31권1호
• /
• pp.1-16
• /
• 2002

Since changepoint identification is important in many data analysis problem, we wish to make inference about the locations of one or more changepoints of the sequence. We consider the Bayesian nonparameteric inference for multiple changepoint problem using a Bayesian segmentation procedure proposed by Yang and Kuo (2000). A mixture of products of Dirichlet process is used as a prior distribution. To decide whether there exists a single change or not, our approach depends on nonparametric Bayesian Schwartz information criterion at each step. We discuss how to choose the precision parameter (total mass parameter) in nonparametric setting and show that the discreteness of the Dirichlet process prior can ha17e a large effect on the nonparametric Bayesian Schwartz information criterion and leads to conclusions that are very different results from reasonable parametric model. One example is proposed to show this effect.

• 응용통계연구
• /
• 제31권3호
• /
• pp.343-352
• /
• 2018

설명변수가 주어졌을 때 반응변수의 평균적인 추세뿐만 아니라 극단적인 지역에서의 추세에 대해서 추정하고 싶거나 반응변수 분포의 일반적인 탐색을 위해서는 분위수 회귀분석과 평률 회귀분석을 사용할 수 있다. 본 논문에서는 평률 회귀모형의 추정을 위한 모수적 방법과 비모수적 방법의 성능을 비교하고자 한다. 이를 위해 각 추정 방법을 소개하고 여러 상황의 모의실험 및 실제자료에의 적용을 통해 비교 분석을 실시하였다. 모형에 따라 성능 차이가 있는데 자료의 형태가 복잡하여 변수 간의 관계를 유추하기 힘들 경우 비모수적으로 추정한 평률 회귀분석모형이 더욱 좋은 결과를 보였다. 일반적인 회귀분석의 경우와 달리 평률의 경우 후보가 되는 모수 모형을 상정하기 어렵다는 측면에서 볼 때, 비모수적 방법의 사용이 추천될 수 있다.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.1443-1448
• /
• 2003

Model Predictive Control (MPC) has recently found wide acceptance in industrial applications, but its potential has been much impounded by linear models due to the lack of a similarly accepted nonlinear modelling or data based technique. The authors have recently developed a new method for obtaining Volterra kernels of up to third order by use of pseudorandom M-sequence. By use of this method, nonparametric NMPC is derived in discrete-time using multi-dimensional convolution between plant data and Volterra kernel measurements. This approach is applied to an industrial polymerisation process using Volterra kernels of up to the third order. Results show that the nonparametric approach is very efficient and effective and considerably outperforms existing methods, while retaining the original data-based spirit and characteristics of linear MPC.