Process manufacturing 2 artificial intelligence
Data
This package comprises business and data understanding. Project requirements and objectives are established from a business perspective and then translated into technical objectives and a project plan. An initial data collection and exploration is done, with the objective of establishing a first contact with the problem. This phase is often critical.
Standardisation
This phase involves selecting, cleaning and generating correct data sets, organised and ready for the modelling phase.
Modelling
Creation of knowledge models from the data supplied from the previous phase.
Training
A segmentation of the available data is carried out for training and subsequent testing and validation. Additionally, it is evaluated whether it is necessary to feed the data using the data augmentation technique. During training, the hyperparameters that define the industrial process under study are obtained for each model chosen.
Testing
After the training of the models, the models obtained are evaluated, taking into account the fulfilment of the success criteria of the business problem posed at the beginning of the project and using the knowledge of the problem domain.
Validation
The chosen model is validated. The knowledge obtained is transformed into actions within the business process, using the models built and the findings obtained in the business production activity.