Abstract:Given a time series S = ((x(1), y(1)), (x(2), y(2)), ...) and a prescribed error bound epsilon, the piecewise linear approximation (PLA) problem with max-error guarantees is to construct a piecewise linear function f such that vertical bar f(x(i))-y(i)vertical bar <= epsilon for all i. In addition, we would like to have an online algorithm that takes the time series as the records arrive in a streaming fashion, and outputs the pieces of f on-the-fly. This problem has applications wherever time series data is being continuously collected, but the data collection device has limited local buffer space and communication bandwidth, so that the data has to be compressed and sent back during the collection process.Prior work addressed two versions of the problem, where either f consists of disjoint segments, or f is required to be a continuous piecewise linear function. In both cases, existing algorithms can produce a function f that has the minimum number of pieces while meeting the prescribed error bound e. However, we observe that neither minimizes the true representation size of f, i.e., the number of parameters required to represent f. In this paper, we design an online algorithm that generates the optimal PLA in terms of representation size while meeting the prescribed max-error guarantee. Our experiments on many real-world data sets show that our algorithm can reduce the representation size of f by around 15% on average compared with the current best methods, while still requiring O(1) processing time per data record and small space.

A Piecewise Linear Approximation Based on a Statistical Model.

Adaptive Estimation Of Time-Varying Parameters In Linear Systems

Dynamic Behavior of Piecewise Linear Approximations

Piecewise Linear Approximations with a Hypercube Partition

Asynchronous Lebesgue Approximation Model for Distributed Continuous-Time Nonlinear Systems

Piecewise Linear Approximation and Its Application in Model Predictive Control

Piecewise Statistic Approximation Based Similarity Measure for Time Series

Error-free Approximation of Explicit Linear MPC Through Lattice Piecewise Affine Expression

Optimal Piecewise Polynomial Approximation for Minimum Computing Cost by Using Constrained Least Squares

Piecewise Linear Approximation of Streaming Time Series Data with Max-Error Guarantees

Piecewise-linear and bilinear approaches to nonlinear differential equations approximation problem of computational structural mechanics

Linearly Constrained Global Optimization Via Piecewise-Linear Approximation

A Statistical Perspective of Sampling Scores for Linear Regression.

Empirical-Likelihood-Based Inference for Partially Linear Models

Model Averaging for Generalized Linear Model with Covariates that are Missing completely at Random

Finite-Time High-Probability Bounds for Polyak–Ruppert Averaged Iterates of Linear Stochastic Approximation

A Nonprobabilistic Approach for Linear Model Selection

Smoothed State Estimation via Efficient Solution of Linear Equations

Constructing Approximations to Bivariate Piecewise-Smooth Functions

Nonlinear Observer Design Using Piecewise-Linear Functions Based Approximations

Constructive PWL Neural Network Approximation