segment | R Documentation |
Performs a fast segmentation algorithm based on the cyber t test and the t statistics. This is a special version for log-ratios or I/NI calls that are assumed to be centered around 0. For segmentation of data with different characteristics you can a) substract the mean/median/mode from your data or b) use the more general version of this algorithm in the R Bioconductor package "fastseg".
segment(x, alpha = 0.05, segMedianT = NULL, minSeg = 3, eps = 0,
delta = 20, maxInt = 40, cyberWeight = 50)
x |
Values to be segmented. |
alpha |
Real value between 0 and 1 is interpreted as the percentage of total points that are considered as initial breakpoints. An integer greater than 1 is interpreted as number of initial breakpoints. Default = 0.05. |
segMedianT |
Vector of length 2. Thresholds on the segment's median. Segments' medians above the first element are considered as gains and below the second value as losses. If set to NULL the segmentation algorithm tries to determine the thresholds itself. If set to 0 the gain and loss segments are not merged. (Default = NULL). |
minSeg |
Minimum length of segments. Default = 3. |
eps |
Real value greater or equal zero. A breakpoint is only possible between to consecutive values of x that have a distance of at least "eps". Default = 0. |
delta |
Positive integer. A parameter to make the segmentation more efficient. If the statistics of a breakpoint lowers while extending the window, the algorithm extends the windows by "delta" more points until it stops. Default = 20. |
maxInt |
The maximum length of a segment left of the breakpoint and right of the breakpoint that is considered. Default = 40. |
cyberWeight |
The "nu" parameter of the cyber t-test. Default = 50. |
A data frame containing the segments.
Guenter Klambauer klambauer@bioinf.jku.at
x <- rnorm(n=500,sd=0.5)
x[150:200] <- rnorm(n=51,mean=3,sd=0.5)
segment(x)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.