• The Korean Journal of Applied Statistics
• /
• v.32 no.4
• /
• pp.607-618
• /
• 2019

A new and interesting task in statistics is to effectively analyze functional data that frequently comes from advances in modern science and technology in areas such as meteorology and biomedical sciences. Functional linear regression with scalar response is a popular functional data analysis technique and it is often a common problem to determine a functional association if a functional predictor variable affects the scalar response in the models. Recently, Kong et al. (Journal of Nonparametric Statistics, 28, 813-838, 2016) established classical testing methods for this based on functional principal component analysis (of the functional predictor), that is, the resulting eigenfunctions (as a basis). However, the eigenbasis functions are not generally suitable for regression purpose because they are only concerned with the variability of the functional predictor, not the functional association of interest in testing problems. Additionally, eigenfunctions are to be estimated from data so that estimation errors might be involved in the performance of testing procedures. To circumvent these issues, we propose a testing method based on fixed basis such as B-splines and show that it works well via simulations. It is also illustrated via simulated and real data examples that the proposed testing method provides more effective and intuitive results due to the localization properties of B-splines.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.10
• /
• pp.4143-4145
• /
• 2014

Background: This study used receiver operating characteristic curve to analyze Surveillance, Epidemiology and End Results (SEER) Ewing sarcoma (ES) outcome data. The aim of this study was to identify and optimize ES-specific survival prediction models and sources of survival disparities. Materials and Methods: This study analyzed socio-economic, staging and treatment factors available in the SEER database for ES. 1844 patients diagnosed between 1973-2009 were used for this study. For the risk modeling, each factor was fitted by a Generalized Linear Model to predict the outcome (bone and joint specific death, yes/no). The area under the receiver operating characteristic curve (ROC) was computed. Similar strata were combined to construct the most parsimonious models. Results: The mean follow up time (S.D.) was 74.48 (89.66) months. 36% of the patients were female. The mean (S.D.) age was 18.7 (12) years. The SEER staging has the highest ROC (S.D.) area of 0.616 (0.032) among the factors tested. We simplified the 4-layered risk levels (local, regional, distant, un-staged) to a simpler non-metastatic (I and II) versus metastatic (III) versus un-staged model. The ROC area (S.D.) of the 3-tiered model was 0.612 (0.008). Several other biologic factors were also predictive of ES-specific survival, but not the socio-economic factors tested here. Conclusions: ROC analysis measured and optimized the performance of ES survival prediction models. Optimized models will provide a more efficient way to stratify patients for clinical trials.

• Korean Chemical Engineering Research
• /
• v.53 no.2
• /
• pp.236-242
• /
• 2015

We present two data-driven modeling methods, partial least square (PLS) and artificial neural network (ANN), to predict the major operating and performance variables of a polymer electrolyte membrane (PEM) fuel cell stack. PLS and ANN models were constructed using the experimental data obtained from the testing of a 30 kW-class PEM fuel cell stack, and then were compared with each other in terms of their prediction and computational performances. To reduce the complexity of the models, we combined a variables importance on PLS projection (VIP) as a variable selection method into the modeling procedure in which the predictor variables are selected from a set of input operation variables. The modeling results showed that the ANN models outperformed the PLS models in predicting the average cell voltage and cathode outlet temperature of the fuel cell stack. However, the PLS models also offered satisfactory prediction performances although they can only capture linear correlations between the predictor and output variables. Depending on the degree of modeling accuracy and speed, both ANN and PLS models can be employed for performance predictions, offline and online optimizations, controls, and fault diagnoses in the field of PEM fuel cell designs and operations.

• Journal of the Korean Data and Information Science Society
• /
• v.22 no.5
• /
• pp.961-966
• /
• 2011

Circular variables are those that have a period in its range. Their examples include direction of animal migration, and time of drug administration, just to mention a few. Statistical analysis of circular variables is quite different from that of linear variable due to its periodic nature. In this paper, the author proposes new circular regression models using geodesic lines on the surface of the sample space of the response and the predictor variables.

• Proceedings of the IEEK Conference
• /
• 2005.11a
• /
• pp.963-966
• /
• 2005

Previous works related to texture coordinate coding of the three-dimensional(3-D) mesh models employed the same predictor as the geometry coder. However, discontinuities in the texture coordinates cause unreasonable prediction. Especially, discontinuities become more serious for the 3-D mesh model with a non-atlas texture image. In this paper, we propose a new coding scheme to remove discontinuities in the texture coordinates by reallocating texture segments according to a coding order. Experiment results show that the proposed coding scheme outperforms the MPEG-4 3DMC standard in terms of compression efficiency. The proposed scheme not only overcome the discontinuity problem by regenerating a texture image, but also improve coding efficiency of texture coordinate compression.

• International Journal of Contents
• /
• v.9 no.4
• /
• pp.11-15
• /
• 2013

"Protein Folding Problem" is considered to be one of the "Great Challenges of Computer Science" and prediction of disordered protein is an important part of the protein folding problem. Machine learning models can predict the disordered structure of protein based on its characteristic of "learning from examples". Among many machine learning models, we investigate the possibility of multilayer perceptron (MLP) as the predictor of protein disorder. The investigation includes a single hidden layer MLP, multi hidden layer MLP and the hierarchical structure of MLP. Also, the target node cost function which deals with imbalanced data is used as training criteria of MLPs. Based on the investigation results, we insist that MLP should have deep architectures for performance improvement of protein disorder prediction.

• Journal of the Korean Statistical Society
• /
• v.35 no.1
• /
• pp.105-114
• /
• 2006

In this paper we propose a simple and computationally attractive difference-based variance estimator in nonparametric regression models with multivariate predictors. We show that the estimator achieves $n^{-1/2}$ rate of convergence for regression functions with only a first derivative when d, the dimension of the predictor, is less than or equal to 4. When d > 4, the rate turns out to be $n^{-4/(d+4)}$ under the first derivative condition for the regression functions. A numerical study suggests that the proposed estimator has a good finite sample performance.

• Journal of Forest and Environmental Science
• /
• v.28 no.2
• /
• pp.63-67
• /
• 2012

Crown ratio is the ratio of live crown length to tree height. It is often used as an important predictor variable for tree growth equation. It indicates tree vigor and is a useful parameter in forest health assessment. The objective of the study was to develop crown ratio prediction models for Tectona grandis. Based on the data set from the temporary sample plots, several non linear equations including logistics, Chapman Richard and exponential functions were tested. These functions were evaluated in terms of coefficient of determination ($R^2$) and standard error of the estimate (SEE). The significance of the estimated parameters was also verified. Plot of residuals against estimated crown ratios were observed. Although the logistic model had the highest $R^2$ and the least SEE, Chapman-Richard and Exponential functions were observed to be more consistent in their predictive ability; and were therefore recommended for predicting crown ratio in the stand.

• Proceedings of the IEEK Conference
• /
• 2006.06a
• /
• pp.745-746
• /
• 2006

In this paper, we propose a new predictive coding scheme for color data of three-dimensional (3-D) mesh models. We exploit connectivity and geometry information to improve coding efficiency. After ordering all vertices in a 3-D mesh model with a vertex traversal technique, we employ a geometry predictor to compress the color data. The predicted color can be acquired by a weighted sum of reconstructed colors for adjacent vertices using both angles and distances between the current vertex and adjacent vertices.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.32B no.11
• /
• pp.1489-1495
• /
• 1995

In this paper, we propose recurrent neural prediction models (RNPM), recurrent neural networks trained as a nonlinear predictor of speech, as a new connectionist model for speech recognition. RNPM modulates its mapping effectively by internal representation, and it requires no time alignment algorithm. Therefore, computational load at the recognition stage is reduced substantially compared with the well known predictive neural networks (PNN), and the size of the required memory is much smaller. And, RNPM does not suffer from the problem of deciding the time varying target function. In the speaker dependent and independent speech recognition experiments under the various conditions, the proposed model was comparable in recognition performance to the PNN, while retaining the above merits that PNN doesn't have.