sdDiff

sdDiff is a function based on the difference operator (or difference order for HALL method) estimating the time-series standard deviation.
The estimation works for time-series generated by Gaussian
random variables with constant standard deviation and multiple changes in mean.
Three estimators are available:<ul>
<li><code>HALL</code> : the so-called HALL-estimator of order 3. For more details see: (1990) Asymptotically optimal difference-based estimation of variance in nonparametric regression. Authors: Hall, Peter and Kay, JW and Titterinton, DM. Biometrika, pages 521--528</li>
<li><code>MAD</code> : the median absolute deviation estimator computed on <code>diff(x)/sqrt(2)</code> with x the vector of datapoints</li>
<li><code>SD</code> : the standard deviation estimator (function sd) computed on <code>diff(x)/sqrt(2)</code> with x the vector of datapoints</li>
</ul>

Penalized parametric change-point detection by functional pruning dynamic programming algorithm. The successive means are constrained using a graph structure with edges defining the nature of the changes These changes can be unconstrained (type std), up or down constrained (type up and down) or constrained by a minimal size jump (type abs). The type null means that the graph allows us to stay on the same segment. To each edge we can associate some additional properties: a minimal gap size, a penalty, some robust parameters (K,a) for biweight (K) and Huber losses (K and a). The user can also constrain the inferred means to lie between some minimal and maximal values. Data is modeled by a cost with possible use of a robust loss, biweight and Huber (see edge parameters K and a). These costs should have a quadratic, log-linear or a log-log representation. This includes quadratic Gaussian cost (type = 'mean'), log-linear cost (type = 'variance', 'poisson' or 'exp') and log-log cost (type = 'negbin'). More details in the paper published in the Journal of Statistical Software: <doi:10.18637/jss.v106.i06>.

Vincent Runge

gfpop

Graph-Constrained Functional Pruning Optimal Partitioning

Toby Hocking

Guillem Rigaill

Daniel Grose

Gaetano Romano

Fatemeh Afghah

Paul Fearnhead

Michel Koskas

Arnaud Liehrmann

sdDiff function

<dl><dt>x</dt>
<dd>vector of datapoints</dd>
<dt>method</dt>
<dd>Three available methods: <code>"HALL"</code>, <code>"MAD"</code> and <code>"SD"</code></dd></dl>

Arguments

sdDiff — sdDiff

<dl>

<dt>x</dt>
<dd>vector of datapoints</dd>


<dt>method</dt>
<dd>Three available methods: <code>"HALL"</code>, <code>"MAD"</code> and <code>"SD"</code></dd>

</dl>

sdDiff: sdDiff

Description

Usage

Value

Arguments

Examples