A Hybrid Deep Learning Model to Forecast Particulate Matter Concentration Levels in Seoul, South Korea

Title : A Hybrid Deep Learning Model to Forecast Particulate Matter Concentration Levels in Seoul, South Korea
Journal : Atmosphere
Authors : Guang Yang, HwaMin Lee, Giyeol Lee
These authors contributed equally to this work : Guang Yang, HwaMin Lee
Corresponding author : Giyeol Lee
DOI : https://doi.org/10.3390/atmos11040348

Abstract
Both long- and short-term exposure to high concentrations of airborne particulate matter (PM) severely affect human health. Many countries now regulate PM concentrations. Early-warning systems based on PM concentration levels are urgently required to allow countermeasures to reduce harm and loss. Previous studies sought to establish accurate, efficient predictive models. Many machine-learning methods are used for air pollution forecasting. The long short-term memory and gated recurrent unit methods, typical deep-learning methods, reliably predict PM levels with some limitations. In this paper, the authors proposed novel hybrid models to combine the strength of two types of deep learning methods. Moreover, the authors compare hybrid deep-learning methods (convolutional neural network (CNN)—long short-term memory (LSTM) and CNN—gated recurrent unit (GRU)) with several stand-alone methods (LSTM, GRU) in terms of predicting PM concentrations in 39 stations in Seoul. Hourly air pollution data and meteorological data from January 2015 to December 2018 was used for these training models. The results of the experiment confirmed that the proposed prediction model could predict the PM concentrations for the next 7 days. Hybrid models outperformed single models in five areas selected randomly with the lowest root mean square error (RMSE) and mean absolute error (MAE) values for both PM10 and PM2.5. The error rate for PM10 prediction in Gangnam with RMSE is 1.688, and MAE is 1.161. For hybrid models, the CNN–GRU better-predicted PM10 for all stations selected, while the CNN–LSTM model performed better on predicting PM2.5.

Acknowledgments
This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT and Future Planning (NRF-2017R1A2B4010570) and Soonchunhyang University Research Fund.

CNN and RNN models for predicting particulate matter in South Korea

Title : CNN and RNN models for predicting particulate matter in South Korea
Published in : The 11th International Conference on Internet (ICONI 2019)
Author : Guang Yang, Thanongsak Xayasouk, HwaMin Lee
Corresponding author : HwaMin Lee
Location : Lotte Hotels & Resorts Hanoi, Hanoi, Vietnam

Abstract
Particulate matter proved to have severe effects on human health. To implement a real-time predicting system for is indispensable for every country face air pollution problem. Based on the previous study, predicting models for local areas are implemented. In this paper, a convolutional network, a recurrent neural network and a hybrid model combines the convolutional layers with the recurrent layers apply to predict future one weeks PM10 and PM2.5 concentration in seven urban cities in South Korea. The main contribution in this paper is: proposed three predicting models, make a comparison of three different models for predicting the concentration of particulate matter, all models are proved capable to obtain reliable predicting results. Experiments show GRU archive best performance for most cities both for PM10 and PM2.5, CNN models speed up the training process with least time consumption. The hybrid model performance stable with less time for training.

Acknowledgments
This research supported by the MSIT(Ministry of Science and ICT), Korea, under the ITRC(Information Technology Research Center) support program(IITP-2019-2015-0-00403) supervised by the IITP(Institute for Information & communications Technology Planning & Evaluation) and Basic Science Research Program through the National Research Foundation of Korea (NRF-2017R1A2B4010570).

실종자 웹 탐색 시스템

제목 : 크롤링을 이용한 자동매칭 게임톡 웹 서비스
저자 : 천태희, 채민수, 양광, 사야속 타농싹, 이화민
게재지 : 2019년 한국정보처리학회 추계학술발표대회 논문집
교신저자 : 이화민
장소 : 제주도, 제주대학교
DOI : https://doi.org/10.3745/PKIPS.y2019m10a.66

초록
실종자는 계속 늘어나는 추세이다. 경찰에서의 실종자 찾는 인력이 부족한 상황이다. 특히 청소년 실종과 달리 성인 실종의 경우 위치정보 파악과 같은 지원을 받기 힘들다. 또한 성인 실종의 경우 단순 가출의 경우가 많아 범죄 가능성이 없을 경우 실종자 탐색에 있어 후순위로 밀린다. 그에 따라 본 논문은 실종자 웹 탐색 시스템을 개발하여 실종자 가족은 실종자를 등록할 수 있게 제공해주며, 실종자정보를 손쉽게 공유할 수 있도록 하였다. 또한 목격자는 실종자를 목격했을 경우 제보하기를 통해서 목격된 정보를 추가할 수 있도록 하였다. 목격자가 제보를 할 경우에는 해당 실종자 가족에게 연락을 가도록 구현하였다.

Acknowledgments
본 연구는 과학기술정보통신부 및 정보통신기기획평가원의 대학ICT연구센터육성지원사업의 연구결과로 수행되었음(IITP-2019-2015-0-00403 & IITP-2019-2014-1-00720)

A hybrid CNN-RNN model for particulate matter forecasting in Seoul

Title : A hybrid CNN-RNN model for particulate matter forecasting in Seoul
Published in : The 14th Asia Pacific International Conference on Information Science and Technology (APIC-IST 2019)
Author : Guang Yang, Thanongsak Xayasouk, HwaMin Lee
Corresponding author : HwaMin Lee
Location : Grand Gongda JianGuo Hotel, Beijing, China

Abstract
Epidemiological research shows that short-term exposure to air pollution especially particulate matter can damage human health. A reliable predicting model with high accuracy is required. Contribute to building real-time forecasting and alarm system. Previous papers focus on improving accuracy by optimizing a single statistical predicting method. In this paper, a hybrid deep learning model is proposed to predict the future day’s particulate matter concentration in Seoul, South Korea. The proposed model combined RNN with a pre-trained ConvNet to forecast future PM2.5 and PM10 concentrations. CNN used to extract the hidden features and used RNN consisted of the gated recurrent unit to make a prediction. With combine the meteorological data, the proposed model can achieve satisfying accuracy and hold less time to consume compared with the bare RNN network.

Acknowledgments
This research supported by the MSIT(Ministry of Science and ICT), Korea, under the ITRC(Information Technology Research Center) support program (IITP-2019-2015-0-00403) supervised by the IITP(Institute for Information & communications Technology Planning & Evaluation) and Basic Science Research Program through the National Research Foundation of Korea (NRF-2017R1A2B4010570).

Fine Dust Predicting using Recurrent Neural Network with GRU

Title : Fine Dust Predicting using Recurrent Neural Network with GRU
Journal : International Journal of Innovative Technology and Exploring Engineering
Authors : Thanongsak Xayasouk, Guang Yang, HwaMin Lee
Corresponding author : HwaMin Lee

Abstract
The particulate matter especially PM2.5 can cause respiratory, cardiovascular and nervous system damage as many studies prove. The monitoring and forecasting system are highly required. This paper proposed a predicting model to forecast PM10 and PM2.5 concentrations in Seoul, South Korea. The proposed model combines the recurrent neural network with GRU. The proposed model can extract the hidden patterns in the long sequence data as RNN’s feature. The proposed model proved they could make satisfying particulate matter concentration in the urban area. The prediction results are reliable even for future 20 days. Meteorological data also contribute to higher predicting results as auxiliary data for the neural network. In further work, we will try to evaluate the model’s universality with more urban cities. Additionally, try to combine other deep learning methods to improve accuracy and reduce time-consuming for prediction.

Acknowledgments
This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (NRF-2017R1A2B4010570) and Soonchunhyang University Research Fund.

Particulate Matter Prediction Using LSTM and GRU

Title : Particulate Matter Prediction Using LSTM and GRU
Published in : The 2019 World Congress on Information Technology Applications and Services (World IT Congress 2019)
Author : Guang Yang, Thanongsak Xayasouk, HwaMin Lee
Corresponding author : HwaMin Lee
Location : Jeju National University, Jeju, Korea

Abstract
Particulate matters (PM) proved can cause cardiovascular, respiratory and nervous system damage, especially PM2.5. Many types of research contributed to making high accuracy prediction of PM concentrations. Machine learning as one of a powerful tool for prediction, have many types of research make a prediction with machine learning models. In this paper, we implemented a prediction system to predict several future days PM10 and PM2.5 concentration with a recurrent neural network (RNN). Consisted of LSTM and GRU units. We predicted several main areas of South Korea and the experiment shows the proposed models can give more than 10 days prediction with high accuracy.

Acknowledgments
This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (NRF-2017R1A2B4010570).

IRCITS 2019 Best Paper Certificate

Title of Paper : Resource-efficient web application framework using Docker
Author : Minsu Chae, Guang Yang, HwaMin Lee
2019.01.25
International Research Confence on Innovation, Technology and Sustainability 2019(IRCITS 2019) Best Paper Certificate

Particulate Matter forecasting using RNN with LSTM

Title : Particulate Matter forecasting using RNN with LSTM
Published in : International Research Conference on Innovation, Technology and Sustainability (IRCITS 2019)
Author : Guang Yang, Minsu Chae, HwaMin Lee
Corresponding author : HwaMin Lee
Location : Century Park Hotel, Manila, Philippines

Abstract
The particulate matter especially PM2.5 can cause respiratory, cardiovascular and nervous system damage as many studies prove. With the increasing concentration of PM in many countries, the prediction of PM became an important issue. In this paper, we tried to create ann recurrent neural network (RNN) with long short-term memory(LSTM) to predict the concentrations of PM2.5 and PM10 7 days in advance in several cities and areas in South Korea.

Acknowledgments
This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning(NRF-2017R1A2B4010570) and the Ministry of Science and ICT, Korea, under the ITRC(Information Technology Research Center) support program (IITP-2018-2014-0-00720) supervised by the IITP(Institute for Information & communications Technology Promotion).

Resource-efficient web application framework using Docker

Title : Resource-efficient web application framework using Docker
Published in : International Research Conference on Innovation, Technology and Sustainability (IRCITS 2019)
Author : Minsu Chae, Guang Yang, HwaMin Lee
Corresponding author : HwaMin Lee
Location : Century Park Hotel, Manila, Philippines

Abstract
Cloud server virtualization technology has become important because of the spread of smart devices, the use of big data, and the proliferation of the Internet of things around the world. Use cloud computing to allocate and deploy IT resources. In case of Docker, performance degradation occurs by using API when monitoring resources. Therefore, in this paper, we propose a framework to manage the container resources according to the user to reduce the load on the Docker API request in the web application.

Acknowledgments
This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning(NRF-2017R1A2B4010570) and the Ministry of Science and ICT, Korea, under the ITRC(Information Technology Research Center) support program (IITP-2018-2015-0-00403) supervised by the IITP(Institute for Information & communications Technology Promotion).

Stacked Autoencoders Model for Fine Particulate Matter (PM10, PM2.5) Prediction

Title : Stacked Autoencoders Model for Fine Particulate Matter (PM10, PM2.5) Prediction
Published in : The 10th International Conference on Internet (ICONI 2018)
Author : Thanongsak Xayasouk, Guang Yang, HwaMin Lee
Corresponding author : HwaMin Lee
Location : Sokha Phnom Penh Hotel, Phnom Penh, Cambodia

Abstract
Nowadays, Fine Particulate matter (PM10, PM2.5) is very harmful to peoples’ health assuming that we know the information about the concentration of fine particulate matter which dangerous for people, it was able to prevent harmful of the fine particulate matter immediately. Deep learning is a type of machine learning method has drawn a lot of academic and industrial interest. This paper presents a deep learning modeling approach, we mostly use the concentrations of particulate matter (PM10, PM2.5) data and meteorological elements in South Korea. We propose a prediction system based on stacked autoencoder model for learning and training data our data. The experimental results, based on Fine Particulate matter and meteorological data, show that the proposed method can achieve great performance in terms of prediction accuracy.

Acknowledgments
This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning(NRF-2017R1A2B4010570) and the Ministry of Science and ICT, Korea, under the ITRC(Information Technology Research Center) support program (IITP-2018-2014-0-00720) supervised by the IITP(Institute for Information & communications Technology Promotion).