This repo contains the source code for paper Zhang L, Ma X, Shi P, et al. RegCNN: A Deep Multi-output Regression Method for Wastewater Treatment[C]//2019 IEEE 31st International Conference on Tools with Artificial Intelligence (ICTAI). IEEE, 2019: 816-823.
Python files with keyword 'prep' are used to preprocess raw data.
table_connect.py joins the data from different tables according to their time stamps.
cnn.py is the training code for RegCNN.
Due to confidentiality issues, only processed data 20200406data_0_ok.txt is provided here.
Recommended tensorflow version is 1.12.
Take cnn.py as an example,
set line 37 'istrain = True', then
python cnn.pyThe training results will be recorded at 'wastewater/res/cnn_res.txt'.
Take cnn.py as an example,
modify line 37: 'istrain = False', then
python cnn.py