• Asia-Pacific Journal of Business
• /
• v.13 no.2
• /
• pp.23-37
• /
• 2022

Purpose - The purpose of this paper is to analyze the dynamic factors of the daily FX implied volatility based on the fractional integration methods focusing on long memory feature and structural changes. Design/methodology/approach - This paper uses the daily FX implied volatility data of the EUR-USD and the JPY-USD exchange rates. For the fractional integration analysis, this paper first applies the basic ARFIMA-FIGARCH model and the Local Whittle method to explore the long memory feature in the implied volatility series. Then, this paper employs the Adaptive-ARFIMA-Adaptive-FIGARCH model with a flexible Fourier form to allow for the structural changes with the long memory feature in the implied volatility series. Findings - This paper finds statistical evidence of the long memory feature in the first two moments of the implied volatility series. And, this paper shows that the structural changes appear to be an important factor and that neglecting the structural changes may lead to an upward bias in the long memory feature of the implied volatility series. Research implications or Originality - The implied volatility has widely been believed to be the market's best forecast regarding the future volatility in FX markets, and modeling the evolution of the implied volatility is quite important as it has clear implications for the behavior of the exchange rates in FX markets. The Adaptive-ARFIMA-Adaptive-FIGARCH model could be an excellent description for the FX implied volatility series

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.6
• /
• pp.881-891
• /
• 2023

Recently, various studies on malicious URL detection using artificial intelligence have been conducted, and most of the research have shown great detection performance. However, not only does classical machine learning require a process of analyzing features, but the detection performance of a trained model also depends on the data analyst's ability. In this paper, we propose a DL-ML Fusion Hybrid Model for malicious web site URL detection based on URL lexical features. the propose model combines the automatic feature extraction layer of deep learning and classical machine learning to improve the feature engineering issue. 60,000 malicious and normal URLs were collected for the experiment and the results showed 23.98%p performance improvement in maximum. In addition, it was possible to train a model in an efficient way with the automation of feature engineering.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.12C
• /
• pp.1686-1691
• /
• 2004

In this paper, we present an improved method to calculate Euclidean distance transform based on Guan's method. Compared to the conventional method, Euclidean distance can be computed faster using Guan's method when the number of feature pixels is small; however, overall computational cost increases proportional to the number of feature pixels in an image. To overcome this problem, we divide feature pixels into two groups: boundary feature pixels (BFPs) and non-boundary feature pixels (NFPs). Here BFPs are defined as those in the 4-neighborhood of foreground pixels. Then, only BFPs are used to calculate the Voronoi diagram resulting in a Euclidean distance map. Experimental results indicate that the proposed method takes 40 Percent less computing time on average than Guan's method. To prove the performance of the proposed method, the computing time of Euclidean distance map by proposed method is compared with the computing time of Guan's method in 16 images that are binary and the size of 512${\times}$512.

• Journal of the Korean Society of Radiology
• /
• v.6 no.3
• /
• pp.167-171
• /
• 2012

In this paper, we present an algorithm to extract features about disease region in digital stomach images. For feature extraction, DCT coefficients of gastrointestinal imaging matrix was obtained. DCT coefficent matrix is concentrated energy in low frequency region, we were extracted 128 feature parameters in low frequency region. Extracted feature parameters can using for differential compression of PACS and, can using for input parameter in CAD.

• ETRI Journal
• /
• v.18 no.1
• /
• pp.29-41
• /
• 1996

This paper describes a method to select a suitable feature for speech recognition using information theoretic measure. Conventional speech recognition systems heuristically choose a portion of frequency components, cepstrum, mel-cepstrum, energy, and their time differences of speech waveforms as their speech features. However, these systems never have good performance if the selected features are not suitable for speech recognition. Since the recognition rate is the only performance measure of speech recognition system, it is hard to judge how suitable the selected feature is. To solve this problem, it is essential to analyze the feature itself, and measure how good the feature itself is. Good speech features should contain all of the class-related information and as small amount of the class-irrelevant variation as possible. In this paper, we suggest a method to measure the class-related information and the amount of the class-irrelevant variation based on the Shannon's information theory. Using this method, we compare the mel-scaled FFT, cepstrum, mel-cepstrum, and wavelet features of the TIMIT speech data. The result shows that, among these features, the mel-scaled FFT is the best feature for speech recognition based on the proposed measure.

• Journal of Institute of Control, Robotics and Systems
• /
• v.2 no.2
• /
• pp.108-114
• /
• 1996

We present proofs of the stability and convergence of Self-organizing feature map (SOFM) neural network with time-invarient learning rate and binary reinforcement function. One of the major problems in Self-organizing feature map neural network concerns with learning rate-"Kalman Filter" gain in stochsatic control field which is monotone decreasing function and converges to 0 for satisfying minimum variance property. In this paper, we show that the stability and convergence of Self-organizing feature map neural network with time-invariant learning rate. The analysis of the proposed algorithm shows that the stability and convergence is guranteed with exponentially stable and weak convergence properties as well.s as well.

• Proceedings of the KIEE Conference
• /
• 2004.11c
• /
• pp.325-327
• /
• 2004

ECG limb lead II signal widely used to diagnosis heart diseases and it is essential to detect ECG events (onsets, offsets and peaks of the QRS complex P wave and T wave) and extract them from ECG signal for heart diseases diagnoses. However, it is very difficult to develop standardized feature extraction formulas since ECG signals are varying on patients and disease types. In this paper, simple feature extraction method from normal and abnormal types of ECG signals is proposed. As a signal features, heart rate, PR interval, QRS interval, QT interval, interval between S wave and baseline, and T wave types are extracted. To show the validity of proposed method, Right Bundle Branch Block (RBBB), Left Bundle Branch Block (LBBB), Sinus Bradycardia, and Sinus Tachycardia data from MIT-BIH arrhythmia database are used for feature extraction and the extraction results showed higher extraction capability compare to conventional formula based extraction method.

• Korean Journal of English Language and Linguistics
• /
• v.1 no.4
• /
• pp.561-585
• /
• 2001

It has been claimed recently that the two computational operations Move and Agree of Chomsky (2000, 2001a) should be separated into distinct and independent operations. According to this view, Move is an “operation that applies only to meet an EPP-feature of a functional category.” It is also claimed under that analysis that “a candidate for Move is simply a syntactic object with phonetic content.” The purpose of this short paper is to show that the operation Move should still be viewed as composite; hence it must have the operation Agree as one of its prerequisites. We argue that the EPP feature of T may not be analyzed as an independent feature that triggers overt displacement in syntax. Under Chomsky's (2000, 2001a) theory, displacement in syntax must require the probe-goal (P, G) association before the actual movement takes place. It is shown in this paper that in order for an element $\beta$ to raise to the [Spec, T] position, the $\varphi$-features of T must establish a (P, G) relation with those of $\beta$ prior to movement. In short, Move requires Agree, the EPP feature being dependent on the minimal $\varphi$-feature [person] of nominals.

• Proceedings of the Korean Society of Tribologists and Lubrication Engineers Conference
• /
• 1998.04a
• /
• pp.314-320
• /
• 1998

Wear debris morphology is closely related to the wear mode and mechanism occured. Image recognition of wear debris is, therefore, a powerful tool in wear monitoring. But it has usually required expert's experience and the results could be too subjective. Development of automatic tools for wear debris recognition is needed to solve this problem. In this work, an algorithm for automatic wear debris recognition was suggested and implemented by PC base software. The presented method defined a characteristic 3-dimensional feature space where typical types of wear debris were separately located by the knowledge-based system and compared the similarity of object wear debris concerned. The 3-dimensional feature space was obtained from multiple feature vectors by using a multi-dimensional scaling technique. The results showed that the presented automatic wear debris recognition was satisfactory in many cases application.

• Journal of the Korean Society for Precision Engineering
• /
• v.13 no.9
• /
• pp.84-93
• /
• 1996

This paper describes the development of new feature positioning method which embedded into the top-down assembly modeling system supporting conceptual design. In this work, the user provides the geometric constraints representing the position and size of features, then the system calculates their proper solution. The use of geometric constraints which are easy to understand intuitively enables the user to represent his design intents about geometric shapes, and enables the system to propagate the changes automatically when some editing occurs. To find the proper solution of given constraints, the Selective Solving Method in which the redundant or conflict equations are detected and discarded is devised. The validity of feature shapes satisfying the constraints can be maintained by this technique, and under or over constrained user-defined constraints can also be estimated. The problems such as getting the initial guess, controlling the multiple solutions, and dealing with objects of rotational symmetry are also resolved. Through this work, the feature based modeling system can support more general and convenient modeling method, and keeps the model being valid during modifying models.