Particle filters are about 25 years old. Initially confined to the so-called “filtering problem” (the sequential analysis of state-space models), they are now routinely applied to a large variety of sequential and non-sequential tasks and have evolved to the broader Sequential Monte Carlo (SMC) framework. This greater applicability comes at the price of a greater technicality. To make matters worse, literature on particle filters spans several scientific fields, mainly engineering (in particular signal processing), but also statistics, machine learning, probability theory, operations research, physics and econometrics. Of course, each field uses slightly different notations and terms to describe the same algorithms. As a result, tracking this literature has become a challenge for non-experts.
An introduction to Sequential Monte Carlo
Papaspiliopoulos, Omiros
2020
Abstract
Particle filters are about 25 years old. Initially confined to the so-called “filtering problem” (the sequential analysis of state-space models), they are now routinely applied to a large variety of sequential and non-sequential tasks and have evolved to the broader Sequential Monte Carlo (SMC) framework. This greater applicability comes at the price of a greater technicality. To make matters worse, literature on particle filters spans several scientific fields, mainly engineering (in particular signal processing), but also statistics, machine learning, probability theory, operations research, physics and econometrics. Of course, each field uses slightly different notations and terms to describe the same algorithms. As a result, tracking this literature has become a challenge for non-experts.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
