멤버십 함수와 DNN을 이용한 PM10 예보 성능의 향상 = Improvement of PM10 Forecasting Performance using Membership Function and DNN

다국어 초록 (Multilingual Abstract)

In this study, we developed a PM10 forecasting model using DNN and Membership Function, and improved the forecasting performance. The model predicts the PM10 concentrations of the next 3 days in the Seoul area by using the weather and air quality observation data and forecast data. The best model(RM14)’s accuracy (82%, 76%, 69%) and false alarm rate(FAR:14%,33%,44%) are good. Probability of detection (POD: 79%, 50%, 53%), however, are not good performance. These are due to the lack of training data for high concentration PM10 compared to low concentration. In addition, the model dose not reflect seasonal factors closely related to the generation of high concentration PM10. To improve this, we propose Julian date membership function as inputs of the PM10 forecasting model. The function express a given date in 12 factors to reflect seasonal characteristics closely related to high concentration PM10.
As a result, the accuracy (79%, 70%, 66%) and FAR (24%, 48%, 46%) are slightly reduced in performance, but the POD (79%, 75%, 71%) are up to 25% improved compared with those of the RM14 model. Hence, this shows that the proposed Julian forecast model is effective for high concentration PM10 forecasts.

참고문헌 (Reference)

1 배효식, "대기질 예보의 성능 향상을 위한 커널 삼중대각 희소행렬을 이용한 고속 자료동화" 한국멀티미디어학회 20 (20): 363-370, 2017

2 Y. Bengio, "Representation Learning : A Review and New Perspectives" 35 (35): 1798-1828, 2013

3 M. Cai, "Prediction of Hourly Air Pollutant Concentrations Near Urban Arterials using Artificial Neural Network Approach" 14 (14): 32-41, 2009

4 A. B. Chelani, "Prediction of Ambient PM10 and Toxic Metals using Artificial Neural Networks" 52 (52): 805-810, 2002

5 W. Lu, "Potential Assessment of A Neural Network Model with PCA/RBF Approach for Forecasting Pollutant Trends in Mong Kok Urban Air, Hong Kong" 96 (96): 79-87, 2004

6 A. Chaloulakou, "Neural Network and Multiple Regression Models for PM10 Prediction in Athens : A Comparative Assessment" 53 (53): 1183-1190, 2003

7 I. G. McKendry, "Evaluation of Artificial Neural Networks for Fine Particulate Pollution (PM10and PM2.5) Forecasting" 52 (52): 1096-1101, 2002

8 J. Schmidhuber, "Deep Learning in Neural Networks: An Overview" 61 (61): 85-117, 2015

9 X. Feng, "Artificial Neural Networks Forecasting of PM2. 5 Pollution Using Air Mass Trajectory Based Geographic Model and Wavelet Transformation" 107 : 118-128, 2015

10 G. Grivas, "Artificial Neural Network Models for Prediction of PM10Hourly Concentrations, in The Greater Area of Athens, Greece" 40 (40): 1216-1229, 2006

11 G. Corani, "Air Quality Prediction in Milan : Feed-Forward Neural Networks, Pruned Neural Networks and Lazy Learning" 185 (185): 513-529, 2005

12 NIER, "A Study of Data Accuracy Improvement for National Air Quality Forecasting(I II )" NIER 2016

13 NIER, "A Study of Construction of Air Quality Forecasting System using Artificial Intelligence(I)" NIER 2017

14 J. Fan, "A Spatiotemporal Prediction Framework for Air Pollution Based on Deep RNN" IV-4/W2 : 15-22, 2017