• The Korean Journal of Applied Statistics
• /
• v.36 no.2
• /
• pp.129-139
• /
• 2023

Forecasting volatility is essential to avoiding the risk caused by the uncertainties of an financial asset. Complicated financial volatility features such as ambiguity between non-stationarity and stationarity, asymmetry, long-memory, sudden fairly large values like outliers bring great challenges to volatility forecasts. In order to address such complicated features implicity, we consider machine leaning models such as LSTM (1997) and GRU (2014), which are known to be suitable for existing time series forecasting. However, there are the problems of vanishing gradients, of enormous amount of computation, and of a huge memory. To solve these problems, a causal temporal convolutional network (TCN) model, an advanced form of 1D CNN, is also applied. It is confirmed that the overall forecasting power of TCN model is higher than that of the RNN models in forecasting VIX, VXD, and VXN, the daily volatility indices of S&P 500, DJIA, Nasdaq, respectively.

• Journal of Internet Computing and Services
• /
• v.24 no.2
• /
• pp.11-18
• /
• 2023

Recently, in the field of Speech Emotion Recognition (SER), many studies have been conducted to improve accuracy using voice features and modeling. In addition to modeling studies to improve the accuracy of existing voice emotion recognition, various studies using voice features are being conducted. This paper, voice files are separated by time interval in a time series method, focusing on the fact that voice emotions are related to time flow. After voice file separation, we propose a model for classifying emotions of speech data by extracting speech features Mel, Chroma, zero-crossing rate (ZCR), root mean square (RMS), and mel-frequency cepstrum coefficients (MFCC) and applying them to a recurrent neural network model used for sequential data processing. As proposed method, voice features were extracted from all files using 'librosa' library and applied to neural network models. The experimental method compared and analyzed the performance of models of recurrent neural network (RNN), long short-term memory (LSTM) and gated recurrent unit (GRU) using the Interactive emotional dyadic motion capture Interactive Emotional Dyadic Motion Capture (IEMOCAP) english dataset.

• Journal of Korea Trade
• /
• v.27 no.1
• /
• pp.79-100
• /
• 2023

Purpose - The purpose of this study is to analyze the market power of the Korea Container Shipping Market (Intra Asia, Korea-Europe, and Korea-U.S.) to verify the existence of collusion empirically, and to answer whether the joint actions of liner market participants in Korea have formed market dominance for each route. Precisely, it will be verified through the Lerner index as to whether the regional market of Asia is a monopoly, oligopoly, or perfect competition. Design/methodology - This study used a Lerner index adjusted with elasticity presented in the New Imperial Organization (NEIO) studies. NEIO refers to a series of empirical studies that estimate parameters to judge market power from industrial data. This study uses B-L empirical models by Bresnahan (1982) and Lau (1982). In addition, NEIO research data statistically contain self-regression and stability problems as price and time series data. A dynamic model following Steen and Salvanes' Error Correction Model was used to solve this problem. Findings - The empirical results are as follows. First, λ, representing market power, is nearly zero in all three markets. Second, the Korean shipping market shows low demand elasticity on average. Nevertheless, the markup is low, a characteristic that is difficult to see in other industries. Third, the Korean shipping market generally remains close to perfect competition from 2014 to 2022, but extreme market power appears in a specific period, such as COVID-19. Fourth, there was no market power in the Intra Asia market from 2008 to 2014. Originality/value - Doubts about perfect competition in the liner market continued, but there were few empirical cases. This paper confirmed that the Korea liner market is a perfect competition market. This paper is the first to implement dynamics using ECM and recursive regression to demonstrate market power in the Korean liner market by dividing the shipping market into Deep Sea and Intra Asia separately. It is also the first to prove the most controversial problems in the current shipping industry numerically and academically.

• The Bulletin of The Korean Astronomical Society
• /
• v.45 no.1
• /
• pp.46.3-47
• /
• 2020

Variations in the brightness of asteroids are caused by their spins, irregular shapes and companions. Thus, in principle, the spin state and shape model of a single object or, a combined model of spins, shapes and mutual orbit of a multiple components can be constructed from the analysis of light curves obtained from the time-series photometry. Using ground- and space-based facilities, a number of time-series photometric observations of asteroids have been conducted to find the possible causes of their light variations. Nonetheless, only about 2% of the known asteroids have been confirmed for their rotation periods. Therefore, a follow-on systematic photometric survey of asteroids is essential. We started an asteroid light curve survey for this purpose using Korea Microlensing Telescope Network (KMTNet) during 199 nights between the second half of 2019 and the first half of 2020. We monitored within a 2° × 14° region of the sky per each night with 25 min cadences. In order to observe as many asteroids as possible with a single exposure, we mostly focus on the ecliptic plane. In our survey, 25,925 asteroids were observed and about 8,000 of them were confirmed for their rotation periods. In addition, using KMTNet's 24-hour continuous monitoring, we collected many composite light curves of slow rotating asteroids that were rarely obtained with previous observations. In this presentation, we will introduce the typical light curves of asteroids obtained from our survey and present a statistical analysis of spin states and shapes of the asteroids from this study.

• International conference on construction engineering and project management
• /
• 2020.12a
• /
• pp.63-74
• /
• 2020

The assembly of modular construction requires a series of thoroughly-considered decisions for crane lifting including the crane model selection, crane location planning, and lift path planning. Traditionally, this decision-making process is empirical and time-consuming, requiring significant human inputs. Recently, research efforts have been dedicated to improving lift planning practices by leveraging cutting-edge technologies such as automated data acquisition, Building Information Modelling (BIM) and computational algorithms. It has been demonstrated that these technologies have advanced lift planning to some degree. However, the advancements tend to be fragmented and isolated. There are two hurdles prevented a systematic improvement of lift planning practices. First, the lack of formalized lift planning workflow, outlining the procedure and necessary information. Secondly, there is also an absence of a shared information environment, enabling storages, updates and the distribution of information to stakeholders in a timely manner. Thus, this paper aims to overcome the hurdles. The study starts with a literature review in combination with document analysis, enabling the initial workflow and information flow. These were contextualised through a series of interviews with Australian practitioners in the crane-related industry, and systematically analysed and schematically validated through an expert panel. Findings included formalized workflow and corresponding information exchanges in a traditional lift planning practice via a Business Process Model and Notation (BPMN). The traditional practice is thus reviewed to identify opportunities for further enhancements. Finally, a BIM-based lift planning workflow is proposed, which integrates the scattered technologies (e.g. BIM and computational algorithms) with the aim of supporting lift planning automation. The resulting framework is setting out procedures that need to be developed and the potential obstacles towards automated lift planning are identified.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.11
• /
• pp.13-20
• /
• 2023

In this paper, we propose a water temperature prediction method using feature extraction and reconstructed data based on LSTM-Autoencoder. We used multivariate time series data such as sea surface water temperature in the Naksan area of the East Sea where the cold water zone phenomenon occurred, and wind direction and wind speed that affect water temperature. Using the LSTM-Autoencoder model, we used three types of data: feature data extracted through dimensionality reduction of the original data combined with multivariate data of the original data, reconstructed data, and original data. The three types of data were trained by the LSTM model to predict sea surface water temperature and evaluated the accuracy. As a result, the sea surface water temperature prediction accuracy using feature extraction of LSTM-Autoencoder confirmed the best performance with MAE 0.3652, RMSE 0.5604, MAPE 3.309%. The result of this study are expected to be able to prevent damage from natural disasters by improving the prediction accuracy of sea surface temperature changes rapidly such as the cold water zone.

• Journal of Internet of Things and Convergence
• /
• v.9 no.6
• /
• pp.111-117
• /
• 2023

There is a lot of research on using a combination of graph neural networks and recurrent neural networks as a way to account for both temporal and spatial dependencies. In particular, graph neural networks are an emerging area of research. Seoul's bicycle rental service (aka Daereungi) has rental stations all over the city of Seoul, and the rental information at each station is a time series that is faithfully recorded. The rental information of each rental station has temporal characteristics that show periodicity over time, and regional characteristics are also thought to have important effects on the rental status. Regional correlations can be well understood using graph neural networks. In this study, we reconstructed the time series data of Seoul's bicycle rental service into a graph and developed a rental prediction model that combines a graph neural network and a recurrent neural network. We considered temporal characteristics such as periodicity over time, regional characteristics, and the degree importance of each rental station.

• Journal of Internet Computing and Services
• /
• v.25 no.2
• /
• pp.93-99
• /
• 2024

In this paper, we derive the Partial Differential Bond Price Equation (PDBPE) by using Ito's Lemma to determine the pricing of bond on term-structure of interest rate (TSIR) model with jump. From PDBPE, the Maclaurin series (MS) and the moment-generating function (MGF) for the exponential function are used to obtain a numerical solution (NS) of the bond prices. And an algorithm for determining bond prices using Monte Carlo Simulation (MCS) techniques is proposed, and the pricing of bond is determined through the simulation process. Comparing the results of the implementation of the above two pricing methods, the relative error (RE) is obtained, which means the ratio of NS and MCS. From the results, we can confirm that the RE is less than around 2.2%, which means that the pricing of bond can be predicted very accurately using the proposed algorithms as well as numerical analysis. Moreover, it was confirmed that the bond price obtained using the MS has a relatively smaller error than the pricing of bond obtained by using the MGF.

• Convergence Security Journal
• /
• v.24 no.2
• /
• pp.31-38
• /
• 2024

Recent expansion in cloud service usage has heightened the importance of cloud security. The purpose of this study is to analyze current research trends in the field of cloud security and to derive implications. To this end, R&D project data provided by the National Science and Technology Knowledge Information Service (NTIS) from 2010 to 2023 was utilized to analyze trends in cloud security research. Fifteen core topics in cloud security research were identified using LDA topic modeling and ARIMA time series analysis. Key areas identified in the research include AI-powered security technologies, privacy and data security, and solving security issues in IoT environments. This highlights the need for research to address security threats that may arise due to the proliferation of cloud technologies and the digital transformation of infrastructure. Based on the derived topics, the field of cloud security was divided into four categories to define a technology reference model, which was improved through expert interviews. This study is expected to guide the future direction of cloud security development and provide important guidelines for future research and investment in academia and industry.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.12 no.3
• /
• pp.93-100
• /
• 2008

The use of fiber reinforced polymer (FRP) composites is significantly growing in construction and infrastructure applications where durability under harsh environmental conditions is of great concern. In order to examine the applicability of FRP rebar as a reinforcement in flexural member, flexural tests were conducted. 12 beams with different FRP materials such as CFRP, GFRP and Hybrid FRP and reinforcement ratio were tested and analyzed in terms of failure mode, moment-deflection, flexural capacity, ductility index and sectional strain distribution. The test results were also compared with the theoretical model represented in ACI 440.1R06. Test results indicate that the flexural capacity of the beams reinforced by FRP bars can be accurately predicted using the ultimate design theory. They also show that the current ACI model for computing the deflection overestimates the actual deflection of GFRP series and underestimates the deflection of CFRP series.