We provide a method to embed knowledge items ( controlling** relationship equations and instance solutions**) from expert experience and experimental results into KDN for time series prediction. Experimental results show that this method greatly improves prediction performance in noisy environments. At the same time, it has higher modeling efficiency than traditional deep learning methods under unstable working conditions in the process industry. We very much look forward to your feedback and will update this technology in a timely manner
e-mail:[email protected]
The warehouse provides part of the testing data set from the actual data set of a steel plant in Jiangyin，China. The training data set requires connection to the database. The user name and password are not provided yet.
If necessary, please contact us
It should be noted that the code of this warehouse has not been fully disclosed.
After the soft copy protection period ends, it will be fully disclosed. Please understand!