• International Journal of Internet, Broadcasting and Communication
• /
• v.15 no.2
• /
• pp.240-247
• /
• 2023

In this paper, effective data management in industrial sites such as intelligent factories using time series data was studied. For effective management of time series data, variables considering the significance of the data were used, and hyper parameters calculated through LSTM AE were applied. We propose an optimized modeling considering the importance of each data section, and through this, outlier data of time series data can be efficiently processed. In the case of applying data significance and applying hyper parameters to which the research in this paper was applied, it was confirmed that the error rate was measured at 5.4%/4.8%/3.3%, and the significance of each data section and the significance of applying hyper parameters to optimize modeling were confirmed.

• 한국데이터정보과학회:학술대회논문집
• /
• 2005.10a
• /
• pp.151-152
• /
• 2005

IGARCH and Stochastic Volatility Model(SVM, for short) have frequently provided useful approximations to the real aspects of financial time series. This article is concerned with modeling various Korean financial time series using both IGARCH and Stochastic Volatility Models. Daily data sets with sample period ranging from 2000 and 2004 including KOSPI, KOSDAQ and won-dollar exchange rate are comparatively analyzed using IGARCH and SVM.

• Journal of the Korean Data and Information Science Society
• /
• v.16 no.4
• /
• pp.835-841
• /
• 2005

IGARCH and Stochastic Volatility Model(SVM, for short) have frequently provided useful approximations to the real aspects of financial time series. This article is concerned with modeling various Korean financial time series using both IGARCH and stochastic volatility models. Daily data sets with sample period ranging from 2000 and 2004 including KOSPI, KOSDAQ and won-dollar exchange rate are comparatively analyzed using IGARCH and SVM.

• Journal of the Korean Data and Information Science Society
• /
• v.16 no.2
• /
• pp.283-288
• /
• 2005

Value at Risk(VaR) has been proven useful in finance literature as a tool of risk management(cf. Jorion(2001)). This article is concerned with introducing VaR to various Korean financial time series. Five daily data sets with sample period ranging from 2000 and 2004 such as KOSPI, KOSPI 200, KOSDAQ, KOSDAQ 50 and won-dollar exchange rate are analyzed using GARCH modeling and in turn VaR is obtained for each data.

• Communications for Statistical Applications and Methods
• /
• v.7 no.1
• /
• pp.87-96
• /
• 2000

Nonlinear time series prediction is derived and compared between statistic of modeling and neural network method. In particular mean squared errors of predication are obtained in generalized random coefficient model and generalized autoregressive conditional heteroscedastic model and compared with them by neural network forecasting.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.22 no.2
• /
• pp.80-92
• /
• 2023

As cities are becoming densely populated, previously unexpected events such as crimes, accidents, and infectious diseases are bound to affect user demands. With a time-series prediction of demand using information with uncertainty, it is impossible to derive reliable results. In particular, the COVID-19 outbreak in early 2020 caused changes in abnormal travel patterns and made it difficult to predict demand for time series. A methodology that accurately predicts demand by detecting and reflecting these changes is, therefore, required. The current study suggests a time series modeling pipeline that automatically detects and predicts abnormal events caused by COVID-19. We expect its wide application in various situations where there is a change in demand due to irregular and abnormal events.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.15 no.1
• /
• pp.1-7
• /
• 2019

The objective of this study is to examine the trends in big data. Research abstracts were extracted from 4,019 articles, published between 1995 and 2018, on Web of Science and were analyzed using topic modeling and time series analysis. The 20 single-term topics that appeared most frequently were as follows: model, technology, algorithm, problem, performance, network, framework, analytics, management, process, value, user, knowledge, dataset, resource, service, cloud, storage, business, and health. The 20 multi-term topics were as follows: sense technology architecture (T10), decision system (T18), classification algorithm (T03), data analytics (T17), system performance (T09), data science (T06), distribution method (T20), service dataset (T19), network communication (T05), customer & business (T16), cloud computing (T02), health care (T14), smart city (T11), patient & disease (T04), privacy & security (T08), research design (T01), social media (T12), student & education (T13), energy consumption (T07), supply chain management (T15). The time series data indicated that the 40 single-term topics and multi-term topics were hot topics. This study provides suggestions for future research.

• Geophysics and Geophysical Exploration
• /
• v.4 no.3
• /
• pp.70-79
• /
• 2001

The integral equation method is a powerful tool for numerical electromagnetic modeling. But the difficulty of this technique is the size of the linear equations, which demands excessive memory and calculation time to invert. This limitation of the integral equation method becomes critical in inverse problem. The conventional Born approximation, where the electric field in the anomalous body is approximated by the background field, is very rapid and easy to compute. However, the technique is inaccurate when the conductivity contrast between the body and the background medium is large. Quasi-linear, quasi-analytical and extended Born approximations are novel approaches to 3-D EM modeling based on the linearization of the integral equations for scattered EM field. These approximation methods are much less time consuming than full integral equation method and more accurate than conventional Born approximation. They we, however, still approximate methods for 3-D EM modeling. Iterative series methods such as modified Born, quasi-linear and quasi-analytical can be used to increase the accuracy of various approximation methods. Comparisons of numerical performance against a full integral equation and various approximation codes show that the iterative series methods are very accurate and almost always converge. Furthermore, they are very fast and easy to implement on a computer. In this study, extended Born series method is developed and it shows more accurate result than that of other series methods. Therefore, Iterative series methods, including extended Born series, open principally new possibilities for fast and accurate 3-D EM modeling and inversion.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.11
• /
• pp.19-27
• /
• 2022

In this study, we propose a language and platform to describe and manage the MLOps(Machine Learning Operations) workflow for time series data anomaly detection. Time series data is collected in many fields, such as IoT sensors, system performance indicators, and user access. In addition, it is used in many applications such as system monitoring and anomaly detection. In order to perform prediction and anomaly detection of time series data, the MLOps platform that can quickly and flexibly apply the analyzed model to the production environment is required. Thus, we developed Python-based AI/ML Modeling Language (AMML) to easily configure and execute MLOps workflows. Python is widely used in data analysis. The proposed MLOps platform can extract and preprocess time series data from various data sources (R-DB, NoSql DB, Log File, etc.) using AMML and predict it through a deep learning model. To verify the applicability of AMML, the workflow for generating a transformer oil temperature prediction deep learning model was configured with AMML and it was confirmed that the training was performed normally.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.19 no.5
• /
• pp.655-660
• /
• 2009

The context prediction algorithms are not suitable to provide real-time personalized service for users in context-awareness environment. The algorithms have problems like time delay in training data processing and the difficulties of implementation in real-time environment. In this paper, we propose a prediction algorithm with user modeling to shorten of processing time and to improve the prediction accuracy in the context prediction algorithm. The algorithm uses moving path of user contexts for context prediction and generates user model by time-series analysis of user's moving path. And that predicts the user context with the user model by sequence matching method. We compared our algorithms with the prediction algorithms by processing time and prediction accuracy. As the result, the prediction accuracy of our algorithm is similar to the prediction algorithms, and processing time is reduced by 40% in real time service environment.