• 대한교통학회지
• /
• 제28권6호
• /
• pp.147-158
• /
• 2010

교통부문 온실가스 저감과 도로의 경쟁력 강화를 위해 교통정보 제공을 통한 수요분산의 관심이 높아지고 있다. 그러나, 이를 위해서는 효율적이며 효과적인 정보제공전략 수립과 운전자 경로전환 행태와 영향요인들에 대한 연구가 선결적으로 필요한 바, 본 연구에서는 도로의 소통상황을 포함한 주행여건과 운전자의 정보매체 선호특성을 고려하여 경로전환 판별모형을 개발하고자 하였다. CART 분석을 이용한 집단구분에서는 주행여건에 따라 3개 군집으로 분류되었으며, 통계적으로 유의하였다. 그리고, CHAID 분석을 통해 경로전환에 영향을 미치는 주행여건과 선호매체 요인들을 통계적으로 유의한 집단으로 구분하여, 경로전환에 영향을 미치는 주요 요인을 파악하였다. 마지막으로, 판별분석을 통해 주행여건과 선호매체가 경로전환에 미치는 영향정도를 파악하고, 경로전환 예측 판별모형식을 개발하였다. 판별모형식 구축 결과, 경로전환은 주행여건에 더 많은 영향을 받는 것으로 나타났으며, 전체 판별적중률(Hit Ratio)은 64.2%로 도출되어 본 판별식은 일정수준 이상의 높은 판별력을 가지고 있었다.

• Communications for Statistical Applications and Methods
• /
• 제9권1호
• /
• pp.87-100
• /
• 2002

Poorly-posed problems in the balanced discriminant analysis was considered. We restrict consideration to the case of observations and the number of variables are the same and small. When these problems exist, we do not use a maximum likelihood estimates(MLE) to estimate covariance matrices. Instead of MLE, an alternative estimate for the covariance matrices are proposed. This alternative method make good use of two regularization parameters, $\lambda$} and $\gamma$. A new test rule for the discriminant function is suggested and examined via limited hut informative simulation study. From the simulation study, it is shown that the suggested test rule gives better test result than other previously suggested method in terms of error rate criterion.

• 한국군사과학기술학회지
• /
• 제14권3호
• /
• pp.467-473
• /
• 2011

The contractor management for the effective defense project is essential factor in the modern defense acquisition task. The occurrence of unlawful company causes hastiness for project manager and setback to the deployment of defense weapon system. In this paper, we develop a prediction model for the effective defense project by using the discriminant analysis and analyse the variables that discriminate the unlawful company in many variables. It is expected that our model can be used to improve the project management capability of defense acquisition and contribute to the establishment of efficient procurement procedure through entry of the reliable defense manufacturer.

• Journal of the Optical Society of Korea
• /
• 제16권1호
• /
• pp.57-62
• /
• 2012

Using continuous wave near-infrared spectroscopy, we measured time-resolved concentration changes of oxy-hemoglobin and deoxy-hemoglobin from the primary motor cortex following finger tapping tasks. These data were processed using partial least squares-discriminant analysis (PLS-DA) to develop a prediction model for a brain-computer interface. The tasks were composed of a series of finger tapping for 15 sec and relaxation for 45 sec. The location of the motor cortex was confirmed by the anti-phasic behavior of the oxy- and deoxy-hemoglobin changes. The results were compared with those obtained using the hidden Markov model (HMM) which has been known to produce the best prediction model. Our data imply that PLS-DA makes better judgments in determining the onset of the events than HMM.

• 말소리와 음성과학
• /
• 제9권2호
• /
• pp.85-94
• /
• 2017

Conventional speaker verification systems using time delay neural network, identity vector and probabilistic linear discriminant analysis (TDNN-Ivector-PLDA) are known to be very effective for verifying long-duration speech utterances. However, when test utterances are of short duration, duration mismatch between enrollment and test utterances significantly degrades the performance of TDNN-Ivector-PLDA systems. To compensate for the I-vector mismatch between long and short utterances, this paper proposes to use probabilistic linear discriminant analysis (PLDA) model adaptation with augmented data. A PLDA model is trained on vast amount of speech data, most of which have long duration. Then, the PLDA model is adapted with the I-vectors obtained from short-utterance data which are augmented by using vocal tract length perturbation (VTLP). In computer experiments using the NIST SRE 2008 database, the proposed method is shown to achieve significantly better performance than the conventional TDNN-Ivector-PLDA systems when there exists duration mismatch between enrollment and test utterances.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권7호
• /
• pp.3194-3216
• /
• 2018

Slow Feature Discriminant Analysis (SFDA) is a supervised feature extraction method inspired by biological mechanism. In this paper, a novel method called Two Dimensional Slow Feature Discriminant Analysis via $L_{2,1}$ norm minimization ($2DSFDA-L_{2,1}$) is proposed. $2DSFDA-L_{2,1}$ integrates $L_{2,1}$ norm regularization and 2D statically uncorrelated constraint to extract discriminant feature. First, $L_{2,1}$ norm regularization can promote the projection matrix row-sparsity, which makes the feature selection and subspace learning simultaneously. Second, uncorrelated features of minimum redundancy are effective for classification. We define 2D statistically uncorrelated model that each row (or column) are independent. Third, we provide a feasible solution by transforming the proposed $L_{2,1}$ nonlinear model into a linear regression type. Additionally, $2DSFDA-L_{2,1}$ is extended to a bilateral projection version called $BSFDA-L_{2,1}$. The advantage of $BSFDA-L_{2,1}$ is that an image can be represented with much less coefficients. Experimental results on three face databases demonstrate that the proposed $2DSFDA-L_{2,1}/BSFDA-L_{2,1}$ can obtain competitive performance.

• 동의생리병리학회지
• /
• 제25권2호
• /
• pp.318-321
• /
• 2011

In this study, when a physician make a diagnosis of the Pattern Identifications(PIs) of stroke patients, the development methods of the PIs classification function is considered by diagnostic questionnaire of the PIs for stroke patients. Clinical data collected from 1,502 stroke patients who was identically diagnosed for the PIs subtypes diagnosed by two clinical experts with more than 3 years experiences in 13 oriental medical hospitals. In order to develop the classification function into PIs using the 44 items-Fire&heat(19), Qi-deficiency(11), Yin-deficiency(7), Dampness phlegm(7)- of them was significant statistically by univariate analysis in 61 questionnaires totally, we make some comparisons of the results of discriminant analysis model and generalized logit model. The overall diagnostic accuracy rate of the PIs subtypes for discriminant model(74.37%) was higher than 3% of generalized logit model(70.09%).

• Geomechanics and Engineering
• /
• 제13권5호
• /
• pp.861-879
• /
• 2017

Coal and gas outburst is a serious dynamic disaster that occurs during coal mining and threatens the lives of coal miners. Currently, coal and gas outburst is commonly predicted using single indicator and its critical value. However, single indicator is unable to fully reflect all of the factors impacting outburst risk and has poor prediction accuracy. Therefore, a more accurate prediction method is necessary. In this work, we first analyzed on-site impacting factors and precursors of coal and gas outburst; then, we constructed a Fisher discriminant analysis (FDA) index system using the gas adsorption index of drilling cutting ${\Delta}h_2$, the drilling cutting weight S, the initial velocity of gas emission from borehole q, the thickness of soft coal h, and the maximum ratio of post-blasting gas emission peak to pre-blasting gas emission $B_{max}$; finally, we studied an FDA-based multiple indicators discriminant model of coal and gas outburst, and applied the discriminant model to predict coal and gas outburst. The results showed that the discriminant model has 100% prediction accuracy, even when some conventional indexes are lower than the warning criteria. The FDA method has a broad application prospects in coal and gas outburst prediction.

• 산업융합연구
• /
• 제15권1호
• /
• pp.17-32
• /
• 2017

The purpose of this study is to evaluate financial models that can predict corporate bankruptcy with diverse studies on evaluation models. The study uses discriminant analysis, logistic model, decision tree, neural networks as analyses tools with 18 input variables as major financial factors. The study found meaningful variables such as current ratio, return on investment, ordinary income to total assets, total debt turn over rate, interest expenses to sales, net working capital to total assets and it also found that prediction performance of suggested method is a bit low compared to that in literature review. It is because the studies in the past uses the data set on the listed companies or companies audited from outside. And this study uses data on the companies whose credibility is not verified enough. Another finding is that models based on decision tree analysis and discriminant analysis showed the highest performance among many bankruptcy forecasting models.

• Communications for Statistical Applications and Methods
• /
• 제12권1호
• /
• pp.125-137
• /
• 2005

Many discriminant analysis models for binary data have been used in real applications, but none of the classification models dominates in all varying circumstances(Asparoukhov & Krzanowski(2001)). Lee and Hwang (2003) proposed a new classification model by using multinomial distribution with the maximum entropy estimation method. The model showed some promising results in case of small number of variables, but its performance was not satisfactory for large number of variables. This paper explores to use the iterative cross entropy minimization estimation method in replace of the maximum entropy estimation. Simulation experiments show that this method can compete with other well known existing classification models.