Multiple Change Points by Filtered Derivative and False Discovery Rate

Let X = (X1; X2; : : : ; Xn) be a time series, that is a sequence of random variable indexed by thetime t = 1; 2; : : : ; n. We assume the existence of a segmentation τ = (τ1; τ2; : : : ; τn) such that Xi is a family of independent identically distributed (i.i.d) random variable for i ∈ (τk; τk + 1]; and k = 0; : : : ; K where by convention τo and τK+1 = N. In the literature, it exist two main kinds of change points detection : The change points on-line and the change points off-line. In this paper, we only look at change point analysis (off-line) when the number of change points is unknown. The obtained result is based on the Filtered Derivative method, with a second step based on the False Discovery Rate. This new method is numerically compared to the Filtered Derivative with p-Value. We also present a real-world application of the Filtered Derivative with False Discovery Rate method (FDqV).

Author (s) Details

Mohamed Elmi
Research Laboratory in Mathematics and Economics (LME), University Djibouti, Republic of Djibouti

View Book  :

Leave a Reply

Your email address will not be published. Required fields are marked *

Previous post The Proof of Correctness of the Hypothesis of Lindelof’s
Next post Research on Semi Supervised Image Segmentation Using Optimal Hierarchical Clustering by Selecting Interested Region as Prior Information