The program is in Tensorflow and python, based in DNN-LSTM is a regression for a ball mill machine This is a regression for the sensor that measures the level of material inside the ball mill, fill level, we used two types of sensors, sound and accelerometer. For the most interested there is a paper with all the information about this project and code
The first step is the creation of the database with the program, 'data_generator.py', then with the programs 'MPL_encoding_vibration.py' and ' MPL_encodig_vibration.py' the dimensional reduction is performed and they obtain the main characteristics of the Data base, and finally with the program 'LSTM_regression.py' the regression is performed.
Everything is done on an HP computer, with Processor: Intel® Core ™ i7-6700K CPU @ 4.00GHz × 8 Menory: 15.6 GiB Graphics: GeForce GTX 980 Ti / PCIe / SSE2 - 6GiB
The whole process takes about 4 hours