Piecewise Aggregation Analysis (PAA)

Applies the Piecewise Aggregation Analysis (PAA) algorithm on a table, treating each column as a single time-series. Splits the time-series into a number of bins and returns the aggregate within each bin. Typical choice of aggregate function is the average.

Definition

Input ports

input table

input

Output ports

output table

output

Configuration

Aggregate (aggregate)

Selects aggregate function to apply to each bin

Output bin number (bin_numbers)

Generates one column for row in the input data containing the bin number that row was given (without padding)

Bins (bins)

Number of output bins or number of samples per bin

Fixed bin size (fixed)

Uses bins of a fixed size instead of a fixed number of bins.

Binning method (inclusion)

Method for determining the value selection for each bin

uneven : allows a different number of value in each bin if input length is not a multiple of bins

pad-first : pads the input data with copies of the first value

pad-last : pads the input data with copies of the last value

overlapping : guarantees that all bins have the same

number of samples but samples may fit multiple bins

Output indices (indices)

Outputs two columns with the start/stop index for each bin. Indices start at zero.Stop index is the first row that is NOT included in the given bin

Examples

• select_intervals.syx

Implementation

class node_paa_sax.PiecewiseAggregationAnalysis