a new representative subset selection and outlier detection method IOS (Isolation forest Outlier detection and Subset selection) has been proposed.IOS can detect outlier and select representative subset of samples simultaneously without y values and reduce prediction errors significantly compared with KS, SPXY and RS methods. The IForest is implemented in R language. All the chemometric methods including IOS for processing datasets are implemented by our research group in MATLAB language.
Install the downloaded packages from local zip or tar.gz file.
To start running this algorithm, load the IsolationForest package through "library(IsolationForestt)" in the R commandline windows
###How to cite:###
Milk data contains one outlier and IOS can be used to detect it and select representative subset.
For any questions, please contact: Zhi-Min Zhang: [email protected]