Say I have a training dataset composed by 128 1-Dimensional time series in form of numpy arrays.

They all correspond to a certain action that I label action_1 and I want to recognize.

What would be the most efficient way to accomplish the following taks:

  • Train a model with the training dataset (the 128 1-D numpy arrays)
  • Ask the model to predict the action in a new test entry (a 1-D numpy array)

This is my very first attempt at Machine Learning, thank you in advance for any indications.


I'm not sure about what's your question exactly. If you are interested in the model architecture time series are the typical example for using RNNs (recurrent neural networks) or some more advanced form of them i.e. LSTM or GRU. They are all implemented in keras (https://keras.io/layers/recurrent/).

You may also can experiment with CNNs (convolutional neural networks). Usually they are used for images but in certain cases they can even outperform RNNs. They are also implemented in keras. (https://keras.io/layers/convolutional/)

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.