LOCI: Local Correlation Integral

In SMLoutliers: Outlier Detection Using Statistical and Machine Learning Methods

Description

We provide an R implementation of the Local Correlation Integral method for detecting outliers as developed by Breunig, et al. (2000), and we follow its description given in Papadimitriou, et al. (2002).

Usage

LOCI(data, alpha)

Arguments

data	Any R data.frame which consists of numeric values only
alpha	a number in the unit interval for the fractional circle search

Details

A simple implementation is provided here. The core function is the distance function. For each observation, a search is made for nearest neighbors within r distance of it, and then for each of these neighbors, we find the number of observations in the fractional circle. Calculations based on multi-granularity deviation factor, MDEF, help in determining the outlier.

Author(s)

Siddharth Jain and Prabhanjan Tattar

References

M.M. Breunig, H.P. Kriegel, R.T. Ng, and J. Sander. Lof: Identifying density-based local outliers. In Proc. SIGMOD Conf., pages 93-104, 2000. Papadimitriou, S., Kitagawa, H., Gibbons, P.B. and Faloutsos, C., 2003, March. Loci: Fast outlier detection using the local correlation integral. In Data Engineering, 2003. Proceedings. 19th International Conference on (pp. 315-326). IEEE.

Examples

data(stiff)
OM <- LOCI(stiff,0.5)
OM

Example output

 [1] "Normal"  "Outlier" "Normal"  "Normal"  "Normal"  "Normal"  "Normal" 
 [8] "Normal"  "Outlier" "Normal"  "Normal"  "Normal"  "Normal"  "Normal" 
[15] "Normal"  "Outlier" "Normal"  "Normal"  "Normal"  "Normal"  "Outlier"
[22] "Normal"  "Normal"  "Normal"  "Normal"  "Normal"  "Normal"  "Normal" 
[29] "Outlier" "Normal" 

SMLoutliers documentation built on May 1, 2019, 9:57 p.m.