I understand that you want to use only one signal as input to a sequence-to-sequence learning model for classification.
Yes, it is possible. In fact, this is a common approach for many sequence classification tasks, such as text classification and speech recognition, you can follow these steps:
- Prepare the data: Convert the error signal into a sequence of tokens.
- Design the model: Choose a sequence-to-sequence learning model architecture, such as a recurrent neural network (RNN) or a long short-term memory (LSTM) network. The model should have an input layer that accepts sequences of tokens and an output layer that predicts the class label.
- Train the model: Train the model on the prepared data.
- Evaluate the model: Evaluate the model on a held-out test set.
Once the model is trained and evaluated, you can use it to classify new error signals into the three classes.
For further information, refer to the documentation links below:
