search

ermshaua / claspy

ClaSPy: A Python package for time series segmentation.
BSD 3-Clause "New" or "Revised" License
75 stars 3 forks source link

[Question] Handle missing data or uneven time intervals #14

Open mheskett opened 1 week ago

mheskett commented 1 week ago

Hi there. Very nice work! I am wondering if theres any way to handle missing data or uneven time intervals. Am I correct to assume that currently you just pass in a 1 Dimensional vector with no indication of which time points were taken? For example in the real world we might be able to take time points at 1minute, 2 minutes, and then 8 minutes, but nothing in between.

ermshaua commented 1 week ago

Hi! Yes, you are correct that ClaSP takes a 1-dimensional vector with time points. The assumption is that the data is recorded in equidistant time steps, e.g. every few milliseconds. If you have data with specific unevenly recorded timestamps, I would suggest that you first synchronise, resample, or impute the data to make it equidistant and then pass it to ClaSP for segmentation.