• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1996.04a
• /
• pp.499-501
• /
• 1996

고로에서 출선된 용선성분중 [Si]은 노열 판단의 대표인자이며, 용선의 열원역할을 함과 동시에 탈P작업의 저해요인이자 slag발생, 전로 slopping발생의 원인이기에 적절한 저[Si] 용선을 제조할 필요가 있다. 또한 [S]는 강의 불순물이자 수요가의 요구로 저[S] 용선제조가 요구된다. 이를위해 후공정인 제강공장에 가기전에 100% 용선예비처리를 실시하고 있다. 본 연구에서는 기존의 용선성분 분석방법으로는 효율적인 용선 예비처리가 불가하므로 시계열분석및 중회귀분석을 통하여 성분을 사전예측하여 시간의 한계를 극복하고, 예측된 성분에 의하여 효율적인 예비처리실시 또는 예비처리 없이 제강직송도 실시하는등 제철소 공정통제 기능의 하나인 용선관제기능의 적극적인 대응이 가능하도록 한다.

• Journal of Korean Society of Transportation
• /
• v.33 no.1
• /
• pp.14-28
• /
• 2015

Numerous research has been conducted using internet search data since the mid-2000s. For example, Google Inc. developed a service predicting influenza patterns using the internet search data. The main objective of this study is to prove the hypothesis that highway traffic indices are similar to the internet search patterns. In order to achieve this objective, a model to predict the number of vehicles entering the expressway and space-mean speed was developed and the goodness-of-fit of the model was assessed. The results revealed several findings. First, it was shown that the Google search traffic was a good predictor for the TCS entering traffic volume model at sites with frequent commute trips, and it had a negative correlation with the TCS entering traffic volume. Second, the Naver search traffic was utilized for the TCS entering traffic volume model at sites with numerous recreational trips, and it was positively correlated with the TCS entering traffic volume. Third, it was uncovered that the VDS speed had a negative relationship with the search traffic on the time series diagram. Lastly, it was concluded that the transfer function noise time series model showed the better goodness-of-fit compared to the other time series model. It is expected that "Big Data" from the internet search data can be extensively applied in the transportation field if the sources of search traffic, time difference and aggregation units are explored in the follow-up studies.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2019.05a
• /
• pp.39-39
• /
• 2019

인구의 도시 집중화로 인하여 다량의 생활용수의 사용에 따라 하천의 자정능력을 초과하여 오염을 유발시키고 있다. 이에 도시하천들의 오염은 점점 심해져 경제적으로 많은 문제를 유발하고 있다. 이러한 하천오염 문제를 과학적으로 대응하기 위해서는 오염물질의 농도 측정 및 데이터 축척을 통한 오염예측이 필수적이라 할 수 있으며, 부산광역시 보건환경정보 공개시스템에서는 하천수질 자동측정망을 설치하여 시간 단위로 오염물질을 측정하고 있다. 그러나 온천천의 하천수질 데이터는 계속 쌓여가고 있는데 이 데이터를 활용해서 하천수질 인자 예측이 거의 이뤄지지 않고 있다. 본 연구에서는 순환신경망 알고리즘을 활용하여 일 단위의 하천수질 인자 예측을 시도하였다. 순환신경망은 인공신경망의 발전된 형태인 시계열 학습에 강한 RNN, LSTM 알고리즘을 활용한 일단위 하천수질 인자 예측을 하고자 하였다. 연구에 앞서 시간 단위로 쌓여있는 데이터를 평균 내어 일 단위로 변경하였고 이 데이터를 가지고 일 단위 하천수질 인자 예측을 진행하였다. 연구에는 Google에서 개발한 딥러닝 오픈소스 라이브러리인 텐서플로우를 활용하여 DO, 탁도 등 항목을 예측하였다. 하천오염의 학습과 예측을 위해 대상지로는 부산지역 온천천의 부곡교, 세병교, 이섭교 관측소를 선택하였다. 연구를 위해 DO, 탁도 등 자료 수집은 부산광역시 보건환경정보 공개시스템의 자료를 활용하였다. 모형의 학습을 위해 입력자료로는 하천수질 인자 자료를 이용하였고, 자료의 학습에는 2014년~2017년 4년간의 자료를 학습자료로 사용하였고, 2018년 1년간의 자료는 모형의 검증을 위해 사용하였다. RNN, LSTM 알고리즘을 활용하여 분석 시 은닉층의 개수, 반복시행횟수, sequence length 등의 값을 조절하여 하천수질 인자 예측을 하였다. 모형의 검증을 위해 $R^2$(r square)와 RMSE(root mean square error)을 이용하여 통계분석을 실시하였다.

• Journal of the Korean Data and Information Science Society
• /
• v.26 no.4
• /
• pp.885-894
• /
• 2015

It is necessary to forecast the electricity demand for reliable and effective operation of the power system. In this study, we try to categorize a functional data, the mean curve in accordance with the time of daily power demand pattern. The data were collected between January 1, 2009 and December 31, 2011. And it were converted to time series data consisting of seasonal components and error component through log transformation and removing trend. Functional clustering by Ma et al. (2006) are applied and parameters are estimated using EM algorithm and generalized cross validation. The number of clusters is determined by classifying holidays or weekdays. Monday, weekday (Tuesday to Friday), Saturday, Sunday or holiday and season are described the mean curve of daily power demand pattern.

• Korean Business Review
• /
• v.16
• /
• pp.43-60
• /
• 2003

There are several stylized facts concerning stock return volatility. First, it is persistent, so an increase in current volatility lasts for many periods. Second, stock volatility increases after stock prices fall. Third, stock volatility is related to macroeconomic volatility, recessions, and banking crises. On the other hand, there are many competing parametric models to represent conditional heteroskedasticity of stock returns. For this article, I adopt the strategy followed by French, Schwert, and Stambaugh(1987) and Schwert(l989, 1990). The models in this article provide a more structured analysis of the time-series properties of stock market volatility. Briefly, these models remove autoregressive and seasonal effects from daily returns to estimate unexpected returns. Then the absolute values of the unexpected returns are used in an autoregressive model to predict stock volatility.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2018.05a
• /
• pp.223-223
• /
• 2018

취수원에서 정수장과 배수지를 거쳐 수용가에 이르기까지 공급되는 급수량을 결정하는데 있어 각 수용가별 물 사용 패턴은 수요량을 예측하여 취수량을 결정하는데 있어 매우 중요한 지표이다. 생활용수 추정은 용도별(가정용, 상업용, 공업용 등)로 분류하여 경향성이 나타날 수 있도록 과거 사용실적을 바탕으로 장래 용도별 사용량을 추정한다. 이는 경험을 바탕으로 한 것으로 일반적으로 시계열 모형을 이용하는데 수요예측의 실패 가능성이 높으며 효율적인 방법이라 할 수 없다. 이에 본 연구에서는 최근 통신기술의 발달로 양방향 통신이 가능한 AMI(Advanced Metering Infrastructure, 원격검침인프라)센서를 영종도 112블록의 528개의 수용가에 설치하였다. AMI는 스마트 미터에서 측정한 데이터를 원격 검침기를 통해 물 사용량을 자동으로 계측할 수 있다. AMI 데이터를 이용하여 영종도 112블록의 운북동과 운서동의 각 용도별, 요일별, 그리고 도심지와 농가의 실시간 물 사용 패턴을 분석하였다. 분석 결과 운북동과 운서동의 물 사용 패턴은 비슷한 경향을 보이는 것으로 보이나 도시화된 운서동에 비해 운북동의 물사용량이 상대적으로 적고 첨두사용량의 발생시간 또한 빠른 것으로 나타났다. 또한 가정용과 공공용의 경우 시간별 물 사용량이 요일에 따라 일정한 경향이 있으나 상업용과 공업용은 일정한 사용량을 보였다. 향후 112블록의 관망해석에 실시간 물사용 패턴을 적용하여 효율적으로 급수량 결정을 할 수 있을 것으로 사료된다.

• Journal of Korea Port Economic Association
• /
• v.39 no.2
• /
• pp.1-13
• /
• 2023

This study forecasted the manpower demand of eco-friendly smart shipbuilding, whose importance and weight are increasing according to the environmental regulations of the IMO and the spread of the 4th industrial revolution technology. It predicted the shipbuilding industry manpower by applying various models of trend analysis and time series analysis based on data from 2000 to 2020 of Statistics Korea. It was found that the prediction applying geometric mean had the smallest gap among the trend and time series analysis methods in comparing between forecast results and actual data for the past 5 years. Therefore, the demand for manpower in the shipbuilding industry was predicted by using the geometric mean method. In addition, the manpower demand of smart eco-friendly ships wast forecasted by using the 2018 and 2020 manpower survey results of the Ministry of Trade, Industry and Energy and reflecting the trend of manpower increase in the shipbuilding industry. The result of forecasting showed that 62,001 person in 2025 and 85,035 people in 2030. This study is expected to contribute to the adjustment of manpower supply and demand and the training professional manpower in the future by increasing the accuracy of forecasting for high value-added eco-friendly smart ships.

• Journal of Digital Convergence
• /
• v.16 no.8
• /
• pp.169-179
• /
• 2018

Recently, web search query information has been applied in advanced predictive model research. Google dominates the global web search market in the Korean market; however, Naver possesses a dominant market share. Based on this characteristic, this study intends to compare the utility of the Korean web search query information of Google and Naver using predictive models. Therefore, this study develops three time-series predictive models to estimate the youth unemployment rate in Korea using the ARIMA model. Model 1 only used the youth unemployment rate in Korea, whereas Models 2 and 3 added the Korean web search query information of Naver and Google, respectively, to Model 1. Compared to the predictability of the models during the training period, Models 2 and 3 showed better fit compared with Model 1. Models 2 and 3 correlated different query information. During predictive periods 1 (continuous with the training period) and 2 (discontinuous with the training period), Model 3 showed the best performance. During predictive period 2, only Model 3 exhibited a significant prediction result. This comparative study contributes to a general understanding of the usefulness of Korean web query information using the Naver and Google search engines.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.24 no.4
• /
• pp.408-414
• /
• 2018

Various studies have been conducted to reduce marine accidents. However, research on marine incidents is only marginal. There are many reports of marine incidents, but the main content of existing studies has been qualitative, which makes quantitative analysis difficult. However, quantitative analysis of marine accidents is necessary to reduce marine incidents. The purpose of this paper is to analyze marine incident data quantitatively by applying big data techniques to predict marine incident trends and reduce marine accident. To accomplish this, about 10,000 marine incident reports were prepared in a unified format through pre-processing. Using this preprocessed data, we first derived major keywords for the Marine incidents at sea using text mining techniques. Secondly, time series and cluster analysis were applied to major keywords. Trends for possible marine incidents were predicted. The results confirmed that it is possible to use quantified data and statistical analysis to address this topic. Also, we have confirmed that it is possible to provide information on preventive measures by grasping objective tendencies for marine incidents that may occur in the future through big data techniques.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.8
• /
• pp.23-30
• /
• 2022

Deep learning analyzes data to discover a series of rules and anticipates the future, helping us in various ways in our lives. For example, prediction of stock prices and agricultural prices. In this research, the results of solar photovoltaic power generation accompanied by weather are analyzed through deep learning in situations where the importance of solar energy use increases, and the amount of power generation is predicted. In this research, we propose a model using LSTM(Long Short Term Memory network) that stand out in time series data prediction. And we compare LSTM's performance with CNN(Convolutional Neural Network), which is used to analyze various dimensions of data, including images, and CNN-LSTM, which combines the two models. The performance of the three models was compared by calculating the MSE, RMSE, R-Squared with the actual value of the solar photovoltaic power generation performance and the predicted value. As a result, it was found that the performance of the LSTM model was the best. Therefor, this research proposes predicting solar photovoltaic power generation using LSTM.