Data Denoising with Weighted Backward Moving Average

Causal Step

This step denoises data similarly to the pre-processing step based on the Backward Moving Average algorithm, but additionally gives a weighting factor to each value within the filter width. Note that the applied algorithm doesn’t need the timestamp information for the denoising procedure.

Input Parameters

1. One-dimensional time series data
2. Filter width
3. Weights for values in the filter width

Output Parameters

1. Denoised time series data

Workflow

Algorithm

Weighted Backward Moving Average

References

• http://demonstrations.wolfram.com/WeightedMovingAverage/