An Event-State Depiction Algorithm Using CPA Methods with Continuous Feed Data

Change Point Analysis (CPA) is a formal method used for determining whether a change-of-state has taken place in a set of observed events. There are claims that this method is capable of detecting subtle changes missed by for example, control charts. By providing confidence levels and confidence intervals it claims to better characterise the changes detected. This is a bonus when for instance, tracking continuous wind velocity data in order to detect when perturbations occur in the oscillation patterns because these are potential predictors of wind gusts, which can have devastating physical effects on land based objects such as buildings and crops. This early-stage research paper describes some previous work in predicting wind gusts using a branch-and-bound algorithm. It considers the need for precision and early detection in wind pattern state-change and examines how fit-for-purpose a change-point analysis method could be for the early detection of velocity oscillation perturbations in a mixed variable analysis of condition change. Wind velocity data is sampled in real-time. The continuous data feed is processed by a change-point analysis algorithm, which has been derived for this purpose. The results are depicted and described.