• Journal of Information Processing Systems
• /
• v.15 no.3
• /
• pp.694-706
• /
• 2019

In this study, we applied the long short-term memory (LSTM) model to classify the cryptocurrency price time series. We collected historic cryptocurrency price time series data and preprocessed them in order to make them clean for use as train and target data. After such preprocessing, the price time series data were systematically encoded into the three-dimensional price tensor representing the past price changes of cryptocurrencies. We also presented our LSTM model structure as well as how to use such price tensor as input data of the LSTM model. In particular, a grid search-based k-fold cross-validation technique was applied to find the most suitable LSTM model parameters. Lastly, through the comparison of the f1-score values, our study showed that the LSTM model outperforms the gradient boosting model, a general machine learning model known to have relatively good prediction performance, for the time series classification of the cryptocurrency price trend. With the LSTM model, we got a performance improvement of about 7% compared to using the GB model.

• Journal of Electrical Engineering and Technology
• /
• v.8 no.3
• /
• pp.544-558
• /
• 2013

This paper proposes a simplified and efficient control scheme for Unified Power Quality Conditioner (UPQC) based on three-level (NPC) inverter capable to mitigate source current harmonics and compensate all voltage disturbances perturbations such us, voltage sags, swells, unbalances and harmonics. The UPQC is designed by the integration of series and shunt active filters (AFs) sharing a common dc bus capacitor. The dc voltage is maintained constant using proportional integral voltage controller. The shunt and series AF are designed using a three-phase three-level (NPC) inverter. The synchronous reference frame (SRF) theory is used to get the reference signals for shunt and the power reactive theory (PQ) for a series APFs. The reference signals for the shunt and series APF are derived from the control algorithm and sensed signals are injected in tow controllers to generate switching signals for series and shunt APFs. The performance of proposed UPQC system is evaluated in terms of power factor correction and mitigation of voltage, current harmonics and all voltage disturbances compensation in three-phase, three-wire power system using MATLAB-Simulink software and SimPowerSystem Toolbox. The simulation results demonstrate that the proposed UPQC system can improve the power quality at the common connection point of the non-linear load.

• Journal of IKEEE
• /
• v.18 no.4
• /
• pp.463-470
• /
• 2014

In stratified flows of two immiscible liquids, due to the vibration in a pipe, the shape of the interface is not always periodic and it causes the different end points of the interfacial boundary. In this case the performance is not good. To solve this, in this paper, the hybrid-type Fourier series is proposed, which consists of both the polynomial and the trigonometric terms. Under the stationary interfacial boundary during acquiring a full set of voltage data, the performance of the proposed method is evaluated through the numerical experiments. The results show that the proposed method performs better than the conventional Fourier series in estimating the interfacial boundary.

• Bulletin of the Korean Chemical Society
• /
• v.33 no.3
• /
• pp.809-817
• /
• 2012

The resonance structure of the preionization spectrum of $H_2$ in the region immediately above its ionization threshold, ($^2{\sum}_{g}^{+}$, $\nu^+=0$, $N^+=0$) converging toward its rotationally excited ($\nu^+=0$, $N^+=2$) limit, is complicated due to perturbation by the vibrationally excited levels $7_{p\pi}\;v=1$ and $57_{p\pi}\;v=2$. The spectra of interlopers are separated from the rotationally preionizing Rydberg series to allow analysis of this complex resonance structure. Although only two vibrationally excited levels perturb the rotational preionization spectrum, at least 6 interloper Rydberg series participate in the complex spectrum over most of its energy range and more interloper series participate at a narrow range around $124500cm^{-1}$ in the spectrum. To allow handling of an arbitrary number of interloper series, MATLAB$^{(R)}$'s symbolic operation is used to perform on-the-fly formulation.

• Proceedings of the Korean Society of Agricultural Engineers Conference
• /
• 2003.10a
• /
• pp.163-166
• /
• 2003

In this study, a grouping of particle-size distributions(PSDs) by means of the fuzzy c-means clustering method(FCM) was presented. The classification was performed with the whole and the major soil series representing pedological origin. In case of the major soil series, PSDs were clustered as $2{\sim}4$ groups and the characteristics of clustering results were quite different between the soil series. It was found that the characteristics of PSDs at center of each class can be explained by formation process of each soil series. In case of whole soil data, PSDs were classified to 8 classes in which 4 classes were single mode and 4 classes were bimode distributions. Through this study, it is concluded that pedogenetic process is a plausible explanation for grain size distribution of soils.

• Research in Mathematical Education
• /
• v.19 no.4
• /
• pp.247-265
• /
• 2015

Mathematics textbook plays a significant role in shaping students' learning of mathematics. Logic, rigor and abstraction as typical features of the formalization of mathematics, dominate mathematics textbooks around the world, which is regarded as one of the important origins of students' learning difficulties in mathematics. An innovative series of Chinese mathematic textbooks is presented in this paper. Supported by the supplementary materials excerpts from the textbooks, it gives a comprehensive theoretical analysis of the principles of design and implementation of this series of mathematics textbooks. The effectiveness of this series of textbooks is demonstrated by student achievement and secondary research data. It shows that series of Chinese mathematic textbooks has largely decreased students' learning difficulties in mathematics and enhance classroom teaching efficiency. It suggests that prioritizing the essence of mathematics and reducing abstraction is an important notion for mathematics textbook design and implementation.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.67 no.2
• /
• pp.209-215
• /
• 2018

This paper present a 7-level PWM inverter adopting voltage balancing control to series-connected input capacitors. The prior proposed 7-level PWM inverter consists of dc input source, three series-connected capacitors, two bidirectional switch modules, and an H-bridge. This circuit topology is useful to increase the number of output voltage levels, however it fails to generate 7-level in output voltage without consideration for voltage balancing among series-connected capacitors. Capacitor voltage imbalance is caused on the different period between charging and discharging of capacitor. To solve this problem, we uses the amplitude modulation of carrier wave, which is used to produce the center output voltage level. To verify the validity of the proposed control method, we carried out computer-aided simulation and experiments using a prototype.

• Journal of Multimedia Information System
• /
• v.8 no.3
• /
• pp.183-190
• /
• 2021

In this paper, we implement a system for a fund recommendation based on the investment propensity and for a future fund price prediction. The investment propensity is classified by scoring user responses to series of questions. The proposed system recommends the funds with a suitable risk rating to the investment propensity of the user. The future fund prices are predicted by Prophet model which is one of the machine learning methods for time series data prediction. Prophet model predicts future fund prices by learning the parameters related to trend changes. The prediction by Prophet model is simple and fast because the temporal dependency for predicting the time-series data can be removed. We implement web pages for the fund recommendation and for the future fund price prediction.

• Korean Journal of Materials Research
• /
• v.3 no.1
• /
• pp.72-83
• /
• 1993

Abstract A new and versatile method for the preparation of biphenyl thioester series was studied in ferroelectric liquid crystal. Many of the resulting liquid crystal showed a ferroelectric chiral smectic C phase in addition to the chiral nematic and smectic A phase. We obtained a large spontaneous polarization exceeding $10^{-8}$C/c$m^2$ in new biphenyl thioester series. It was found that the size and polarity of the substituent in chiral carbon influenced the magnitude of spontaneous polarization. The series allow us to determine the influence of substituent size in the chiral carbon the existence and the stability of the Sc* phase.

• Proceedings of the KIEE Conference
• /
• 1998.07g
• /
• pp.2299-2301
• /
• 1998

In this paper, a modified RBF (Radial Basis Function) neural network structure is suggested for the prediction of time series with non-linear, non-stationary characteristics. Conventional RBF neural network predicting time series by using past outputs is for sensing the trajectory of the time series and for reacting when there exists strong relation between input and hidden neuron's RBF center. But this response is highly sensitive to level and trend of time serieses. In order to overcome such dependencies, hidden neurons are modified to react to the increments of input variable and multiplied by increments(or decrements) of out puts for prediction. When the suggested structure is applied to prediction of Lorenz equation, and Rossler equation, improved performances are obtainable.