Derivative Extrapolation Using Least Squares

TL;DR

This paper introduces three differentiation methods for streaming data, including an optimization of Savitzky-Golay, tested on synthetic and real WIFI data, analyzing their benefits and limitations.

Contribution

It presents novel adaptations of differentiation techniques for streaming data, with a focus on optimizing the Savitzky-Golay method for WIFI data analysis.

Findings

All methods are applicable to streaming data.

The optimized Savitzky-Golay method performs best on WIFI data.

Insights into benefits and pitfalls of each method are provided.

Abstract

Here, we present three methods for differentiating discrete sets from streaming processes, e.g. WIFI. One approach is based on optimization of the well-known Savitzky-Golay algorithm. These methods are tested on synthetic data sets and will be implemented on subsets of a real university campus WIFI data set. The applicability of all methods are discussed, where we provide insights on both some of their benefits and pitfalls. This article ends with our conclusion on which method is better for our WIFI data.