A decision-based multi-layered outlier detection system for resource constraint manet

Abstract

MANET is a useful network for providing various services and applications. Among those services and applications, sharing is important. The sharing of resources is possible when the availability of resources is ensured. In this work, the multi-dimensional multi-layered solution is proposed for ensuring the availability of network resources. The multi-dimensional approach provides criteria for collecting and analyzing data from different security dimensions. A multi-layered outlier detection algorithm using hierarchical data interconnection is proposed in this work. In the analysis, it is observed that internal indices like DBI and RSI give confirmation of clusters stability with the proposed approach. A minimum of 4.1% and a maximum of 11.3% stability is observed with variation in a number of nodes. Similarly, external indices like F-measure and NMI indicate stability in comparison to external clusters. A minimum of 2% and a maximum of 13.5% stability is observed.