• Journal of the Korea Institute of Military Science and Technology
• /
• v.16 no.5
• /
• pp.675-682
• /
• 2013

Reliability analysis and prediction of next failure time is critical to sustain weapon systems, concerning scheduled maintenance, spare parts replacement and maintenance interventions, etc. Since 1981, many methodology derived from various probabilistic and statistical theories has been suggested to do that activity. Nowadays, many A.I. tools have been used to support these predictions. Support Vector Regression(SVR) is a nonlinear regression technique extended from support vector machine. SVR can fit data flexibly and it has a wide variety of applications. This paper utilizes SVM and SVR with combining time series to predict the next failure time based on historical failure data. A numerical case using failure data from the military equipment is presented to demonstrate the performance of the proposed approach. Finally, the proposed approach is proved meaningful to predict next failure point and to estimate instantaneous failure rate and MTBF.

• Journal of Information Technology Services
• /
• v.19 no.4
• /
• pp.49-61
• /
• 2020

Prior studies on technology predictions attempted to predict the emergence and spread of emerging technologies through the analysis of correlations and changes between data using objective data such as patents and research papers. Most of the previous studies predicted future technologies only from the viewpoint of technology development. Therefore, this study intends to conduct technical forecasting from the perspective of the consumer by using keyword search frequency of search portals such as NAVER before and after the introduction of emerging technologies. In this study, we analyzed healthcare technologies into three types : measurement technology, platform technology, and remote service technology. And for the keyword analysis on the healthcare, we converted the classification of technology perspective into the keyword classification of consumer perspective. (Blood pressure and blood sugar, healthcare diagnosis, appointment and prescription, and remote diagnosis and prescription) Naver Trend is used to analyze keyword trends from a consumer perspective. We also used the ARIMA model as a technology prediction model. Analyzing the search frequency (Naver trend) over 44 months, the final ARIMA models that can predict three types of healthcare technology keyword trends were estimated as "ARIMA (1,2,1) (1,0,0)", "ARIMA (0,1,0) (1,0,0)", "ARIMA (1,1,0) (0,0,0)". In addition, it was confirmed that the values predicted by the time series prediction model and the actual values for 44 months were moving in almost similar patterns in all intervals. Therefore, we can confirm that this time series prediction model for healthcare technology is very suitable.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.19 no.5
• /
• pp.595-602
• /
• 2009

Fuzzy rules, which represent the behavior of their system, are sensitive to fuzzy clustering techniques. If the classification abilities of such clustering techniques are improved, their systems can work for the purpose more accurately because the capabilities of the fuzzy rules and parameters are enhanced by the clustering techniques. Thus, this paper proposes a new hierarchically structured clustering algorithm that can enhance the classification abilities. The proposed clustering technique consists of two clusters based on correlationship and statistical characteristics between data, which can perform classification more accurately. In addition, this paper uses difference data sets to reflect the patterns and regularities of the original data clearly, and constructs multiple fuzzy systems to consider various characteristics of the differences suitably. To verify effectiveness of the proposed techniques, this paper applies the constructed fuzzy systems to the field of time series prediction, and performs prediction for nonlinear time series examples.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2001.01a
• /
• pp.122-127
• /
• 2001

This study investigates the effectiveness of time delay neural networks(TDNN) for the time dependent prediction domain. Although it is well-known fact that the back-propagation neural network(BPN) performs well in pattern recognition tasks, the method has some limitations in that it can only learn an input mapping of static (or spatial) patterns that are independent of time of sequences. The preliminary results show that the accuracy of TDNN is higher than the standard BPN with time lag. Our proposed approaches are demonstrated by the stork market prediction domain.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2018.07a
• /
• pp.141-142
• /
• 2018

기존 데이터의 패턴 예측에는 통계를 기반으로 한 수학적 모델이 주로 사용되었으나 새로운 데이터에 대한 피드백이 부족하기 때문에 장기간의 데이터 예측에 한계가 있다. 또한 데이터의 특성이 다양하고 복잡한 경우에는 수학적 모델의 결합 및 계산과정이 어려워진다. 따라서 본 논문에서는 데이터의 학습 및 예측에 기존 정적 모델이 아닌 기계학습 중 시계열 데이터 분석 (Time Series Analysis) 을 기반으로 연구를 진행하였다. 기계학습은 복잡한 특성을 가진 데이터를 학습하여 미래의 데이터 값을 예측하거나 분류하는데 있어서 정확도 및 처리시간 측면에서의 성능을 향상시킬 수 있다.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2000.05a
• /
• pp.224-227
• /
• 2000

In this Paper, we prepose a method of constructing neural networks using bio-inspired emergent and evolutionary concepts. This method is algorithm that is based on the characteristics of the biological DNA and growth of plants. Here is, we propose a constructing method to make a DNA coding method for production rule of L-system. L-system is based on so-called the parallel rewriting mechanism. The DNA coding method has no limitation in expressing the production rule of L-system. Evolutionary algorithms motivated by Darwinian natural selection are population based searching methods and the high performance of which is highly dependent on the representation of solution space. In order to verify the effectiveness of our scheme, we apply it to one step ahead prediction of Mackey-Glass time series, Sun spot data and KOSPI data.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.3
• /
• pp.528-550
• /
• 2024

Warehousing demand prediction is an essential part of the supply chain, providing a fundamental basis for product manufacturing, replenishment, warehouse planning, etc. Existing forecasting methods cannot produce accurate forecasts since warehouse demand is affected by external factors such as holidays and seasons. Some aspects, such as consumer psychology and producer reputation, are challenging to quantify. The data can fluctuate widely or do not show obvious trend cycles. We introduce a new model for warehouse demand prediction called MAGRU, which stands for Multi-layer Attention with GRU. In the model, firstly, we perform the embedding operation on the input sequence to quantify the external influences; after that, we implement an encoder using GRU and the attention mechanism. The hidden state of GRU captures essential time series. In the decoder, we use attention again to select the key hidden states among all-time slices as the data to be fed into the GRU network. Experimental results show that this model has higher accuracy than RNN, LSTM, GRU, Prophet, XGboost, and DARNN. Using mean absolute error (MAE) and symmetric mean absolute percentage error(SMAPE) to evaluate the experimental results, MAGRU's MAE, RMSE, and SMAPE decreased by 7.65%, 10.03%, and 8.87% over GRU-LSTM, the current best model for solving this type of problem.

• Journal of IKEEE
• /
• v.25 no.4
• /
• pp.766-769
• /
• 2021

In order to perform tasks such as design, control, optimization, and prediction of flight vehicle trajectories based on machine learning techniques including deep learning, a certain amount of flight vehicle trajectory data is required. However, there are cases in which it is difficult to secure more than a certain amount of flight vehicle trajectory data for various reasons. In such cases, synthetic data generation could be one way to make machine learning possible. In this paper, to explore this possibility, we generated and evaluated synthetic flight vehicle trajectory data using time-series generative adversarial neural network. In addition, various ablation studies (comparative experiments) were performed to explore the possibility of using synthetic data in the aircraft trajectory prediction task. The experimental results presented in this paper are expected to be of practical help to researchers who want to conduct research on the possibility of using synthetic data in the generation of synthetic flight vehicle trajectory data and the work related to flight vehicle trajectories.

• 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.

• Advanced Industrial SCIence
• /
• v.3 no.1
• /
• pp.17-22
• /
• 2024

History, which began with the emergence of mankind, has a means of recording. Today, we can check the past through data. Generated data may only be generated and stored at a certain moment, but it is not only continuously generated over a certain time interval from the past to the present, but also occurs in the future, so making predictions using it is an important task. In order to find out trends in the use of time series data among numerous data, this paper analyzes the concept of time series data, analyzes Recurrent Neural Network and Long-Short Term Memory, which are mainly used for time series data analysis in the machine learning field, and analyzes the use of these models. Through case studies, it was confirmed that it is being used in various fields such as medical diagnosis, stock price analysis, and climate prediction, and is showing high predictive results. Based on this, we will explore ways to utilize it in the future.