What kind of Neural Network should I build to classify each instance of a time series sequence?

1

Let’s say, I have the time-series dataset below-left. I would like to train a model in such a way that if I feed the model with an input like below-right, it should be able to classify each sample with the correct class label.

    Training Sequence:             Test Sequence:

Time,   Bitrate,   Class           Time,   Bitrate      Predicted Class
 0,       312,       1              0,       234      -----> 0
 0.3,     319,       1              0.2,     261      -----> 0
 0.5,     227,       0              0.4,     277      -----> 0
 0.6,     229,       0              0.7,     301      -----> 1 
 0.7,     219,       0              0.8,     305      -----> 1
 0.8,     341,       1              0.9,     343      -----> 1 
 0.9,     281,       0              1.0,     299      -----> 0 
          ...                                ...

Any help is appreciated. Thanks!

2

Are the columns (time, bitrate) the input features? If so, I guess you could try to train a small classifier or it seems even a threshold such as return bitrate > 300 could perfectly classify the data.

3

This is a network experiment, therefore the only feature column is the bitrate. However, the actual data is not like this, I mean that one simple threshold wouldn’t work. I just wrote those values to demonstrate.

4

That sounds like a sequence labeling task, which is common in NLP, e.g., to assign each word in a sentence a Part-of-Speech tag. You should be able to easily adopt the linked example if it fits the bill.

5

Thank you Chris! I will definitely have a look at this.