How to perform time-series classification for small time-series?

Question

I am trying to perform a time-series classification using features. This means that the feature extraction algorithm calculates characteristics such as the average or maximal value of the time series and use it for classification.

However, my time-series is small have about 10 data points each.

e.g., [10, 10, 10, 10, 8, 5, 3, 2, 1, 1], [8, 8, 8, 8, 9, 5, 4, 3, 2, 1], and so on.

In that case what are the features I can extract?

I came accross a really cool python package named as tsfresh that calculates a huge number of time-series features. However, since my time-series is small I am worried if that is suitable for me.

I am happy to provide more details if needed.

Answer 1

A time series can have 1) short term arima structure ; 2) seasonal arima structure ; 3) seasonal deterministic structure ; 4) one or more level/step shifts ; 5) one or more deternnistic trends ; 6) one or more pulses ; 7) changes in model parameters over time ; 8) changes in model error variance over time.

Are these the kinds of inferential features that you wish to use to classify your data or are you just trying to use basic descriptive statistics like the mean,mediation,range,standard deviation, coefficient of variation et al.