Windowing & Aggregation¶

How do I analyze temporal dynamics?

Features in the cobrabox.windowing domain. Access them as cb.<Feature> (canonical), or as cb.windowing.<Feature> / cb.feature.<Feature>.

Each feature has two forms: a class for building pipelines, and a one-shot function for a single call.

Aggregate a stream of per-window Data by stacking along a new 'window' dimension.

Tags: aggregation, io:preserves-time

Aggregate a stream of per-window Data by averaging across windows.

Tags: aggregation, io:scalar-per-channel

Yield one Data per sliding window over the time dimension.

cb.sliding_window(data, ...)  # one-shot
cb.SlidingWindow(...).apply(data)  # composable, for pipelines

Tags: temporal-dynamics, segmentation, nonstationarity, io:iterator

Sliding window with automatic per-window reduction.

cb.sliding_window_reduce(data, ...)  # one-shot
cb.SlidingWindowReduce(...).apply(data)  # composable, for pipelines

Tags: temporal-dynamics, segmentation, reduction, nonstationarity, io:preserves-time