Browsing by Author "Jou, Yow-Jen"

Now showing 1 - 17 of 17

A new multicollinearity diagnostic for generalized linear models
(2016)
Huang, Chien-Chia L.
;
Jou, Yow-Jen
;
Cho, Hsun-Jung
;
運輸與物流管理系註：原交通所+運管所
;
資訊管理與財務金融系註：原資管所+財金所
We propose a new collinearity diagnostic tool for generalized linear models. The new diagnostic tool is termed the weighted variance inflation factor (WVIF) behaving exactly the same as the traditional variance inflation factor in the context of regression diagnostic, given data matrix normalized. Compared to the use of condition number (CN), WVIF shows more reliable information on how severe the situation is, when data collinearity does exist. An alternative estimator, a by-product of the new diagnostic, outperforms the ridge estimator in the presence of data collinearity in both aspects of WVIF and CN. Evidences are given through analyzing various real-world numerical examples.
A VIF-based optimization model to alleviate collinearity problems in multiple linear regression
(2014-12-01)
Jou, Yow-Jen
;
Huang, Chien-Chia Liam
;
Cho, Hsun-Jung
;
運輸與物流管理系註：原交通所+運管所
;
資訊管理與財務金融系註：原資管所+財金所
In this paper, we address data collinearity problems in multiple linear regression from an optimization perspective. We propose a novel linearly constrained quadratic programming model, based on the concept of the variance inflation factor (VIF). We employ the perturbation method that involves imposing a general symmetric non-diagonal perturbation matrix on the correlation matrix. The proposed VIF-based model reduces the largest VIF by minimizing the resulting biases. The VIF-based model can mitigate the harm from data collinearity through the reduction in both the condition number and VIFs, meanwhile improving the statistical significance. The resulting estimator has bounded biases under an iterative framework and hence is termed the least accumulative bias estimator. Certain potential statistical properties can be further considered as the side constraints for the proposed model. Various numerical examples validate the proposed approach.
Difference-based matrix perturbation method for semi-parametric regression with multicollinearity
(2017-01-01)
Huang, Chien-Chia L.
;
Jou, Yow-Jen
;
Cho, Hsun-Jung
;
運輸與物流管理系註：原交通所+運管所
;
資訊管理與財務金融系註：原資管所+財金所
This paper addresses the collinearity problems in semi-parametric linear models. Under the difference-based settings, we introduce a new diagnostic, the difference-based variance inflation factor (DVIF), for detecting the presence of multicollinearity in semi-parametric models. The DVIF is then used to device a difference-based matrix perturbation method for solving the problem. The electricities distribution data set is analyzed, and numerical evidences validate the effectiveness of the proposed method.
Functional Canonical Analysis Between Functional and Interval Data
(2009)
Jou, Yow-Jen
;
Huang, Chien-Chia
;
Wu, Jennifer Yuh-Jen
;
資訊管理與財務金融系註：原資管所+財金所
;
Department of Information Management and Finance
In this study we discuss the functional canonical correlation analysis between the functional data and the interval data. To address the interval data, a representative is of necessity. Based on the work by Chavent et al. (2002), the representative can be derived by using the Hausdorff distance between intervals. The canonical analysis can be either the mixed functional-multivariate canonical correlation analysis or a pure functional one. This approach is then applied to the roadside vehicle detection by using Radar devices. It can be observed that the weight functions implicitly contain the information about the distances between the specific lane and the detector.
Incomplete information analysis for the origin-destination survey table
(2006-12-01)
Jou, Yow-Jen
;
Cho, Hsun-Jung
;
Lin, Pei-Wei
;
Wang, Chih-Yin
;
統計學研究所
Sampling is one approach used to survey the origin-destination (O-D) trip matrix. However, when the sampling rate is not sufficiently large compared to the population, the sampling data may have missing values in O-D pairs and that makes the O-D matrix incomplete. Two imputation methods to solve the problem mentioned previously are presented in this study. The Deming-Stephan (D-S) proportional fitting procedure is a statistical method first proposed to impute the missing value in this study. The improved D-S method, its convergence is proved by Cauchy criteria in this study, applies the iteration conception in the D-S method to solve the incomplete data problem. One numerical example displayed in this study shows that the improved D-S method produces an estimated O-D table with similar pattern as that of the population table.
INDUSTRY COMPETITION, AGENCY PROBLEM, AND FIRM PERFORMANCE
(2014-01-01)
Wang, Sue-Fung
;
Jou, Yow-Jen
;
Chang, Ke-Chiun
;
Wu, Kun-Wei
;
資訊管理與財務金融系註：原資管所+財金所
The study investigates the relationship between industry competition and firm performance in relation to agency problem conditions. At the same time, the literature also stated that different levels of national corporate governance have different impacts on corporate performance. We also investigate whether or not the relation between industry competition and firm performance would be different under different levels of agency problem. Our empirical results show that the positive relation between industry competition and firm performance is relatively significant for firms with higher free cash flows that present severe agency problems. The results do not change under different levels of national corporate governance.
Method for identification of traffic lane boundary
(2010-06-24)
Jou, Yow-Jen
;
Cho, Hsun-Jung
;
Chen, Yu-Kuang
;
Huang, Heng
;
Hsu, Chia-Chun
The invention provides a method for identification of traffic lane boundary. Firstly the microwave signal is received, and the noise reduction is treated for the microwave signal. Then the frequency domain information is employed to calculate the legal set of closed interval, in order to form the frequency span information. Finally, the probability density function model is employed to calculate the frequency span information in order to identify the traffic lane boundary.
Parallel chain convergence of time dependent origin-destination matrices with gibbs sampler
(2006)
Jou, Yow-Jen
;
Ch, Hsun-Jung
;
Lan, Chien-Lun
;
Hsu, Chia-Chun
;
資訊管理與財務金融系註：原資管所+財金所
An effective method of O-D estimation by the state-space model has been introduced by Jon. Coupled with Gibbs sampler and Kalman filter, the state-space model can generated precious O-D matrices without any prior information while other studies assume that the transition matrix is known or at least approximately known. The Gibbs sampler, a particular type of Markov Chain Monte Carlo method, is one of the iterative simulation methods. To monitor of convergence of this iterative simulation, a parallel chain technique is implemented in this paper. By the numerical example, the convergence of the different chains would be clearly pointed out. The comparison of simulation and real data also shows that satisfying results can be obtained by the model.
Path Flow Estimation Using Time Varying Coefficient State Space Model
(2009)
Jou, Yow-Jen
;
Lan, Chien-Lun
;
資訊管理與財務金融系註：原資管所+財金所
;
Department of Information Management and Finance
The dynamic path flow information is very crucial in the field of transportation operation and management, i.e., dynamic traffic assignment, scheduling plan, and signal timing. Time-dependent path information, which is important in many aspects, is nearly impossible to be obtained. Consequently, researchers have been seeking estimation methods for deriving valuable path flow information from less expensive traffic data, primarily link traffic counts of surveillance systems. This investigation considers a path flow estimation problem involving the time varying coefficient state space model, Gibbs sampler, and Kalman filter. Numerical examples with part of a real network of the Taipei Mass Rapid Transit with real O-D matrices is demonstrated to address the accuracy of proposed model. Results of this study show that this time-varying coefficient state space model is very effective in the estimation of path flow compared to time-invariant model.
Semi-parametric linear mixed effects model for vehicles identification
(2007)
Jou, Yow-Jen
;
Yang, Chai-Tzu
;
Huang, Chien-Chia
;
Wu, Jennifer Yuh-Jen
;
資訊管理與財務金融系註：原資管所+財金所
In order to make the detecting of the lanes and the types of the vehicles traveling on various roadways affordable, radio-frequency (RF) system-on-chip is designed and will be mounted on the roadside to collect vehicle information. We use the Fast Fourier Transform (FFT) to transform the signal of the reflecting wave radar into the numerical data, and utilize it by the statistical approach to discriminate the size of cars and the lanes. In order to classify the types of the vehicles, two models are proposed to model the data. One is multivariate analysis of variance model to account for the main effect and the interaction effect between type and lane, the other is the semi-parametric linear mixed effect model to emphasize the functional characteristic of the data. Both models work well when the number of groups is small but deteriorate when the number of groups increases.
Signal recognition by nonparametric discrimination of functional data for vehicle transportation system
(2007)
Jou, Yow-Jen
;
Lu, Yung-Tsai
;
Lan, Chien-Lun
;
Lee, Hsia
;
資訊管理與財務金融系註：原資管所+財金所
Traffic congestion is a serious and general problem in our daily life. Real time vehicle information is essential to the advanced traffic management. The recognition and statistics of traffic flow among different types of vehicles would be contributive to improve traffic block. This paper considers the dataset recorded by the radar microwave detector with the intensity of reflecting waves and the types of vehicles. The data is treated as functional data and then classification would be proposed to be performed by nonparametfic discrimination of functional data with three forms of Proximity. The proximity with lower misclassification rate would be adopted for the data. The results show that the misclassification rate is pretty satisfactory if the number of groups is two and as the number of group increases the misclassification rate increases as expected.
Statistical queuing-length approach to system-on-chip signal control
(2005)
Jou, Yow-Jen
;
Tseng, Ming-Te
;
統計學研究所
;
Institute of Statistics
The chief goal of this paper is to describe a methodology for predicting the dynamic queue length of traffic signal control system on chip (SoC). There are two key steps. The first is to predict the dynamic traffic volumes for an oversaturated traffic intersection's link. This step is to fit a censored model for the traffic flow by using the historical relevant information of traffic flow, including past volumes and occupancy data. Then, the predicted traffic volumes will be cumulated to predict the queue length. Also, algorithm has been developed and a sample is given to verify the model.
The Online Vehicle Type Classifier Design for Road-Side Radar Detectors
(2009)
Jou, Yow-Jen
;
Chen, Yu-Kuang
;
資訊管理與財務金融系註：原資管所+財金所
;
Department of Information Management and Finance
This study presents an online vehicle type classifier for road-side radar detectors in multi-lane environments. An automatic learning framework which composes a parametric statistic model and algorithms is introduced. The parameters of an online vehicle type classifier are trained with vehicles passing in front of detectors. The online vehicle type classifier tries to identify the vehicle type in real time. The road-side radar detector is developed based on frequency-modulation continuous-wave (FMCW) radar with the carrier frequency at X-band. Vehicles are classified into two major categories, large and small. The classification based on (i) average energy maximum and (ii) average energy variance, that are extracted from the frequency-domain signatures caused by passed vehicles. A two-dimension Gaussian Mixed Model (denoted as GMM) is employed to develop the learning model. Expectation maximization (denoted EM) algorithm is implemented to obtain the parameters of GMM. Numerical examples are demonstrated with real-world experiments. In the field tests, the automatic framework delivers an accuracy of minimum 88%, even with extremes scenarios (including (i) small samples and (ii) large sample size difference of different vehicle types). The examples show satisfying results of the proposed online vehicle type classifier.
Time Dependent Origin-destination Estimation from Traffic Count without Prior Information
(2009-06-01)
Cho, Hsun-Jung
;
Jou, Yow-Jen
;
Lan, Chien-Lun
;
運輸與物流管理系註：原交通所+運管所
;
資訊管理與財務金融系註：原資管所+財金所
Existing research works on time-dependent origin-destination (O-D) estimation focus on the surveillance data usually assume the prior information of the O-D matrix (or transition matrix) is known (or at least partially known). In this paper, we relax such assumption by combining Gibbs sampler and Kalman filter in a state space model. A solution algorithm with parallel chain convergence control is proposed and implemented. To enhance its efficiency, a parallel structure is suggested with efficiency and speedup demonstrated using PC-cluster. Two numerical examples (one for Taipei Mass Rapid Transit network and the other for Taiwan Area National Freeway network) are included to show the proposed model could be effective of time-dependent origin-destination estimation.
User-oriented travel time prediction using grey system and real-time vehicle detector data
(2005)
Jou, Yow-Jen
;
Lan, Chien-Lun
;
統計學研究所
;
Institute of Statistics
Real-time travel information is increasingly important for ITS. Many high-valued time-sensitive applications can take advantage from this real-time information. This paper describes an approach to estimate travel time with Grey System. Combined with real-time vehicle detector data, wireless communication ability and global positioning system, individual driving preference is then measured and been used to adjust the estimation model. An empirical result is shown at the end of this paper and shows a satisfying result.
VIF-based adaptive matrix perturbation method for heteroskedasticity-robust covariance estimators in the presence of multicollinearity
(2017)
Huang, Chien-Chia Liam
;
Jou, Yow-Jen
;
Cho, Hsun-Jung
;
運輸與物流管理系註：原交通所+運管所
;
資訊管理與財務金融系註：原資管所+財金所
In this study, we investigate linear regression having both heteroskedasticity and collinearity problems. We discuss the properties related to the perturbation method. Important observations are summarized as theorems. We then prove the main result that states the heteroskedasticity-robust variances can be improved and that the resulting bias is minimized by using the matrix perturbation method. We analyze a practical example for validation of the method.
VIF-based adaptive matrix perturbation method for heteroskedasticity-robust covariance estimators in the presence of multicollinearity
(2017)
Huang, Chien-Chia Liam
;
Jou, Yow-Jen
;
Cho, Hsun-Jung
;
運輸與物流管理系註：原交通所+運管所
;
資訊管理與財務金融系註：原資管所+財金所
In this study, we investigate linear regression having both heteroskedasticity and collinearity problems. We discuss the properties related to the perturbation method. Important observations are summarized as theorems. We then prove the main result that states the heteroskedasticity-robust variances can be improved and that the resulting bias is minimized by using the matrix perturbation method. We analyze a practical example for validation of the method.