• 한국정보통신설비학회:학술대회논문집
• /
• 2009.08a
• /
• pp.382-386
• /
• 2009

Online or sequential learning is one of the most basic and powerful method to train neuron network, and it has been widely used in disease detection, weather prediction and other realistic classification problem. At present, there are many algorithms in this area, such as MRAN, GAP-RBFN, OS-ELM, SVM and SMC-RBF. Among them, SMC-RBF has the best performance; it has less number of hidden neurons, and best efficiency. However, all the existing algorithms use signal normal distribution as kernel function, which means the output of the kernel function is same at the different direction. In this paper, we use multi-variable normal distribution as kernel function, and derive EKF learning formulas for multi-variable normal distribution kernel function. From the result of the experience, we can deduct that the proposed method has better efficiency performance, and not sensitive to the data sequence.

• 한국정보디스플레이학회:학술대회논문집
• /
• 2003.07a
• /
• pp.673-674
• /
• 2003

The reliability of TFT, crystallized by sequential lateral solidification (SLS) technology, has been studied High current damage is characterized by high gate bias (-20V) and drain bias (-10V). It is found that performance of SLS TFTs is enhanced by high current stress up to 300 sec of stress time for 20/8 (W/L) N-TFT. After that, TFT performance is degraded with the increase of the stress time. It is speculated from the experimental data that SLS TFTs initially contain a number of unstable defect states. Then, the defect states seem to be cured by high current stress.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.27 no.3
• /
• pp.47-53
• /
• 1990

Polygonal approximation of digital curves is useful for the image analysis or data compression. There are methods of polygonal approximation using cone intersection using cone intersection which have relatively smaller number of break points and are executed in sequential process. Here a method of polygonal approximation is proposed, which is modified from Sklansky and Gonzales' method, and improves the speed by using integer operations.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.16 no.7
• /
• pp.1377-1382
• /
• 2012

In this paper, we propose a novel recognition method that extract source data from encoded signal that are displayed on FND mounted on home appliances. First of all, it find a candidate FND region from sequential difference images taken by smartphone and extract segment image using clustering RGB value. After that, it normalize segment images to correct a slant error and recognize each segments using a relative distance. Experiments show the robustness of the recognition algorithm on smartphone.

• The Korean Journal of Applied Statistics
• /
• v.7 no.2
• /
• pp.183-200
• /
• 1994

Data-adaptive sequential suboptimal designs for very complex computer simulation codes are considered based on a spatial prediction model. These designs are constructed for two simulators of the computational nuclear fusion devices model. The difficulty of constructing the optimal designs due to the irregular design region, and its alternatives are also discussed with some computational algorithms for obtaining the designs.

• Computers and Concrete
• /
• v.3 no.5
• /
• pp.313-334
• /
• 2006

A novel formulation aiming to achieve optimal design of reinforced concrete (RC) structures is presented here. Optimal sizing and reinforcing for beam and column members in multi-bay and multistory RC structures incorporates optimal stiffness correlation among all structural members and results in cost savings over typical-practice design solutions. A Nonlinear Programming algorithm searches for a minimum cost solution that satisfies ACI 2005 code requirements for axial and flexural loads. Material and labor costs for forming and placing concrete and steel are incorporated as a function of member size using RS Means 2005 cost data. Successful implementation demonstrates the abilities and performance of MATLAB's (The Mathworks, Inc.) Sequential Quadratic Programming algorithm for the design optimization of RC structures. A number of examples are presented that demonstrate the ability of this formulation to achieve optimal designs.

• Speech Sciences
• /
• v.11 no.4
• /
• pp.75-88
• /
• 2004

In this paper, we propose a new approach to sequential linear regression adaptation of continuous density hidden Markov models (CDHMMs) based on transformation space model (TSM). The proposed TSM which characterizes the a priori knowledge of the training speakers associated with maximum likelihood linear regression (MLLR) matrix parameters is effectively described in terms of the latent variable models. The TSM provides various sources of information such as the correlation information, the prior distribution, and the prior knowledge of the regression parameters that are very useful for rapid adaptation. The quasi-Bayes (QB) estimation algorithm is formulated to incrementally update the hyperparameters of the TSM and regression matrices simultaneously. Experimental results showed that the proposed TSM approach is better than that of the conventional quasi-Bayes linear regression (QBLR) algorithm for a small amount of adaptation data.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2005.05a
• /
• pp.274-282
• /
• 2005

Customer retention is one of the major issued in life insurance industry, in which competition is increasingly fierce. There are many things to do to retain customers. One of those things is to be continuously in touch with all customers. The objective of this study is to design the contact scheduling system(CSS) to support the planers who must touch the customers without having subjective information. Support-planers suffer from lack of information which can be used to intimately touch. CSS that is developed in this study generates contact schedule to touch customers by taking into account existing contact history. CSS has a two stage process. In the first stage, it segments customers according to his or her demographics and contract status data. Then it finds typical pattern and pattern is combined to business rules for each segment. We expert that CSS would support support-planers to make uncontacted customers' experience positive.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2000.11a
• /
• pp.325-332
• /
• 2000

데이터마이닝 분야에서 시계얼 데이터(time-series data)내에서 숨어 있는 순차패턴의 발견은 상품(Items)이나 어떤 사건(Event)과 같이 데이터의 특징이 명확한 대상에 대한 연구는 많이 되어왔으나 수치 값을 가지는 시계열 데이터에서 이들 내부에 숨어 있는 패턴을 발견하는 것은 최근에 관심을 가지게 되었다. 우리는 시계열 데이터를 시간적 변화에 따라 값의 변화 경향(Trend)이 같은 데이터 그룹을 패턴 요소인 벡터 (Vestor)로 표현하여 이들을 이용해서 흥미로운 패턴들을 발견한다. 이와 같은 벡터적인 표현으로 우리는 벡터들 간의 포함관계를 적용해 모든 가능한 형태의 패턴 발견을 목적으로 한다. 또한 경향성을 가진 패턴 요소를 사건(Event)과 같이 취급함으로써 다양한 종류의 시계열 데이터가 동시에 발생될 때 이들 상호간에 연관된 시간적 패턴을 찾을 수 있다. 따라서 이 연구에서 제안하는 경향성을 기초로 한 순차패턴의 탐식은 기업내부의 판매실적의 변화 패턴이나, 고객의 구매 행동분석에 적용이 가능하리라 여겨진다

• Korean Journal of Remote Sensing
• /
• v.19 no.5
• /
• pp.381-392
• /
• 2003

Multi-temporal approaches using sequential data acquired over multiple years are essential for satisfactory discrimination between many land-cover classes whose signatures exhibit seasonal trends. At any particular time, the response of several classes may be indistinguishable. A harmonic model that can represent seasonal variability is characterized by four components: mean level, frequency, phase and amplitude. The trigonometric components of the harmonic function inherently contain temporal information about changes in land-cover characteristics. Using the estimates which are obtained from sequential images through spectral analysis, seasonal periodicity can be incorporates into multi-temporal classification. The Normalized Difference Vegetation Index (NDVI) was computed for one week composites of the Advanced Very High Resolution Radiometer (AVHRR) imagery over the Korean peninsula for 1996 ～ 2000 using a dynamic technique. Land-cover types were then classified both with the estimated harmonic components using an unsupervised classification approach based on a hierarchical clustering algorithm. The results of the classification using the harmonic components show that the new approach is potentially very effective for identifying land-cover types by the analysis of its multi-temporal behavior.