[포스터]Prediction of Water Quality in Water Treatment Plant using Recurrent Neural Network Model |
---|
학술지명 IWA Aspire 2017
저자 한승희,최두용,김경필
발표일 2017-09-11
|
In this paper, we predicted water quality after water treatment using ANN (artificial neural network) and RNN (recurrent neural network) among machine learning algorithms. RNN is mainly used for time series data processing such as natural language processing and speech recognition. RNN is a network model that uses the output of hidden layer for the previous input as the present input of hidden layer, and the past inputs and the states of hidden layer continuously affect the future results. ANN and RNN were built based on 16,000 sets of the two years real water treatment plant data. Then, we predicted the water quality using ANN and RNN with the sets of data for the last three months of the plant, and the two models were compared and analyzed. As a result, it was confirmed that the water quality of the plant can be predicted by two models. ANN was predicted the tendency of water quality, but accuracy was lower than RNN. RNN showed better prediction than ANN. |