• Korean Chemical Engineering Research
• /
• v.59 no.2
• /
• pp.191-199
• /
• 2021

A distillation column, which is a main facility of the chemical process, separates the desired product from a mixture by using the difference of boiling points. The distillation process requires the optimization and the prediction of operation because it consumes much energy. The target process of this study is difficult to operate efficiently because the composition of feed flow is not steady according to the supplier. To deal with this problem, we could develop a data-driven model to predict operating conditions. However, data preprocessing is essential to improve the predictive performance of the model because the raw data contains outlier and noise. In this study, after optimizing the predictive model based long-short term memory (LSTM) and Random forest (RF), we used a low-pass filter and one-class support vector machine for data preprocessing and compared predictive performance according to the method and range of the preprocessing. The performance of the predictive model and the effect of the preprocessing is compared by using R2 and RMSE. In the case of LSTM, R2 increased from 0.791 to 0.977 by 23.5%, and RMSE decreased from 0.132 to 0.029 by 78.0%. In the case of RF, R2 increased from 0.767 to 0.938 by 22.3%, and RMSE decreased from 0.140 to 0.050 by 64.3%.

• The Journal of Asian Finance, Economics and Business
• /
• v.8 no.8
• /
• pp.297-309
• /
• 2021

In this study, to capture the skewness and kurtosis detected in both conditional and unconditional return distributions of the stock markets of Kazakhstan and Russia, two versions of normal mixture GARCH models are employed. The data set consists of daily observations of the Kazakhstan and Russia stock prices, and world crude oil price, covering the period from 1 June 2006 through 1 March 2021. From the empirical results, incorporating the long memory effect on the returns not only provides better descriptions of dynamic behaviors of the stock market prices but also plays a significant role in improving a better understanding of the return dynamics. In addition, normal mixture models for time-varying volatility provide a better fit to the conditional densities than the usual GARCH specifications and has an important advantage that the conditional higher moments are time-varying. This implies that the volatility skews implied by normal mixture models are more likely to exhibit the features of risk and the direction of the information flow is regime-dependent. The findings of this study contain useful information for diverse purposes of cross-border stock market players such as asset allocation, portfolio management, risk management, and market regulations.

• ETRI Journal
• /
• v.35 no.6
• /
• pp.1068-1074
• /
• 2013

This paper proposes an additional forward error correction (FEC) layer to compensate for the defectiveness inherent in the conventional FEC layer in the Long Term Evolution specifications. The proposed additional layer is called a graceful degradation (GD)-FEC layer and maintains desirable service quality even under burst data loss conditions of a few seconds. This paper also proposes a non-delayed decoding (NDD)-GD-FEC layer that is inherent in the decoding process. Computer simulations and device-based tests show a better loss recovery performance with a negligible increase in CPU utilization and occupied memory size.

• Proceedings of the KIEE Conference
• /
• 2004.05a
• /
• pp.147-149
• /
• 2004

Although most of the people have studied English for a long time, their English conversation capability is low. When we provide them portable conversational-type English learners by the application of computer and information process technology, such portable learners can be used to enhance their English conversation capability by their conventional conversation exercises. The core technology to develop such learner is the development of a voice recognition and synthesis module under an embedded environment. This paper deals with voice recognition and synthesis, prototype of the learner module using a DSP(Digital Signal Processing) chip for voice processing, voice playback function, flash memory file system, PC download function using USB ports, English conversation text function by the use of SMC(Smart Media Card) flash memory, LCD display function, MP3 music listening function, etc. Application areas of the prototype equipped with such various functions are vast, i.e. portable language learners, amusement devices, kids toy, control by voice, security by the use of voice, etc.

• Proceedings of the Korean Nuclear Society Conference
• /
• 1996.05a
• /
• pp.349-354
• /
• 1996

SACOM(Simulation Analyser with Cognitive Operator Model) is being developed at Korea Atomic Energy Research Institute to simulate human operator's cognitive characteristics during the emergency situations of nuclear power plans. An operator model with error mechanisms has been developed and combined into SACOM to simulate human operator's cognitive information process based on the Rasmussen's decision ladder model. The operational logic for five different cognitive activities (Agents), operator's attentional control (Controller), short-term memory (Blackboard), and long-term memory (Knowledge Base) have been developed and implemented on blackboard architecture. A trial simulation with a scenario for emergency operation has been performed to verify the operational logic. It was found that the operator model with error mechanisms is suitable for the simulation of operator's cognitive behavior in emergency situation.

• KIPS Transactions on Computer and Communication Systems
• /
• v.11 no.12
• /
• pp.419-428
• /
• 2022

Since in-place updates for pages are not allowed in flash memory, all new page writes should be written in an out-of-place manner. The old overwritten pages are invalidated. Such invalidated pages eventually trigger the costly garbage collection process. Since the garbage collection causes numerous read and write operations, it is one of the flash memory's major performance issues. In 2R, it modified the garbage collection algorithm, which applies the I/O characteristics of the On-Line Transaction Process workload to improve the Write Amplification Factor. However, this algorithm has a region pollution problem. Therefore, in this paper, we developed 2R++ that additionally separates pages with long access intervals to solve the region pollution problem. 2R++ introduces an extra bit per block to separate warm pages based on a second chance mechanism. Prevents warm pages from being misidentified as cold pages to solve region pollution problem. We conducted the experiments on TPC-C and Linkbench to make the performance comparison. The experiment showed that 2R++ achieved a Write Amplification Factor improvement of 57.8% and 13.8% compared to 2R, respectively.

• Smart Media Journal
• /
• v.10 no.1
• /
• pp.63-69
• /
• 2021

When a person sees a sentence and understands the sentence, the person understands the sentence by reminiscent of the main word in the sentence as an image. Text-to-image is what allows computers to do this associative process. The previous deep learning-based text-to-image model extracts text features using Convolutional Neural Network (CNN)-Long Short Term Memory (LSTM) and bi-directional LSTM, and generates an image by inputting it to the GAN. The previous text-to-image model uses basic embedding in text feature extraction, and it takes a long time to train because images are generated using several modules. Therefore, in this research, we propose a method of extracting features by using the attention mechanism, which has improved performance in the natural language processing field, for sentence embedding, and generating an image by inputting the extracted features into the GAN. As a result of the experiment, the inception score was higher than that of the model used in the previous study, and when judged with the naked eye, an image that expresses the features well in the input sentence was created. In addition, even when a long sentence is input, an image that expresses the sentence well was created.

• Journal of the Korean Data and Information Science Society
• /
• v.28 no.3
• /
• pp.697-707
• /
• 2017

In this paper, we propose a modified GARCH(p, q)-X model which is obtained by adding the exogenous variables to the modified GARCH(p, q) process. Some limiting properties are shown under various stationary and nonstationary exogenous processes which are generated by another process independent of the noise process. The proposed model extends the GARCH(1, 1)-X model studied by Han (2015) to various GARCH(p, q)-type models such as GJR GARCH, asymptotic power GARCH and VGARCH combined with exogenous process. In comparison with GARCH(1, 1)-X, we expect that many stylized facts including long memory property of the financial time series can be explained effectively by modified GARCH(p, q) model combined with proper additional covariate.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.52 no.9
• /
• pp.533-539
• /
• 2003

Reinforcement learning (RL) has been widely used as a learning mechanism of an artificial life system. However, RL usually suffers from slow convergence to the optimum state-action sequence or a sequence of stimulus-response (SR) behaviors, and may not correctly work in non-Markov processes. In this paper, first, to cope with slow-convergence problem, if some state-action pairs are considered as disturbance for optimum sequence, then they no to be eliminated in long-term memory (LTM), where such disturbances are found by a shortest path-finding algorithm. This process is shown to let the system get an enhanced learning speed. Second, to partly solve a non-Markov problem, if a stimulus is frequently met in a searching-process, then the stimulus will be classified as a sequential percept for a non-Markov hidden state. And thus, a correct behavior for a non-Markov hidden state can be learned as in a Markov environment. To show the validity of our proposed learning technologies, several simulation result j will be illustrated.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.3A
• /
• pp.337-346
• /
• 2004

Punctured convolutional codes increase transmission efficiency without increasing hardware complexity. However, Viterbi decoder supporting punctured codes requires long decoding length and large survivor memory to achieve sifficiently low bit error rate (BER), when compared to the Viterbi decoder for a rate 1/2 convolutional code. This Paper presents novel architecture adopting a pipelined trace-forward unit reducing survivor memory requirements in the Viterbi decoder. The proposed survivor path architecture reduces the memory requirements by removing the initial decoding delay needed to perform trace-back operation and by accelerating the trace-forward process to identify the survivor path in the Viterbi decoder. Experimental results show that the area of survivor path unit has been reduced by 16％ compared to that of conventional hybrid survivor path unit.