20200521_Water Quality Prediction Method Based on IGRA and LSTM

 

1. Title, Journal, and Authors

Title: Water Quality Prediction Method Based on IGRA and LSTM

Journal: Water

Authors: Jian Zhou1,2,*, Yuanyuan Wang1,2, Fu Xiao1,2, Yunyun Wang1,2, and Lijuan Sun1,2

1College of Computer, Nanjing University of Posts and Telecommunications

2Jiangsu High Technology Research Key Laboratory for Wireless Sensor Networks

 

2. Summary

This paper was written to predict the water quality using a machine learning method with time-series data. Data were pretreated by IGRA, used for making feature selection, learned by LSTM, used for predicting time-series data. Besides, for the comparison of prediction accuracy, the BP network and ARIMA were introduced.

In the results, IGRA showed the more suitable feature selection results, compared with the advanced researches. And, LSTM displayed the lowest RMSE, compared with the others. For progressing the future researches, enough water quality data are required. Added, GPU having high performance, would be needed to decrease the analyzing time.

 

3. Originality and Creativity

IGRA method was introduced in this paper to make feature selection for the LSTM model. Using this method, researchers can select the input parameters optimally.

 

4. Contact

Dae Seong Jeong / Integrated Ph.D. program

 

Environmental Systems Engineering Lab.

School of Environmental Science & Engineering

Gwangju Institute of Science and Technology

1 Oryong-dong Buk-gu Gwangju, 500-712, Korea

 

Phone : +82-10-2003-7860

E-mail : jeongds92@gist.ac.kr