An Optimized Pattern Recognition Algorithm for Anomaly Detection in IoT Environment

TL;DR

This paper presents an optimized string pattern recognition algorithm designed for anomaly detection in IoT environments, utilizing fixed-value mappings and arithmetic operations to improve search accuracy in large datasets.

Contribution

The paper introduces a novel optimized string searching algorithm tailored for anomaly detection in IoT, demonstrating its implementation and testing on large DNA sequence datasets.

Findings

Reduced mismatches in string search results

Effective pattern recognition in large datasets

Identification of algorithm weaknesses

Abstract

With the advent of large-scale heterogeneous search engines comes the problem of unified search control resulting in mismatches that could have otherwise avoided. A mechanism is needed to determine exact patterns in web mining and ubiquitous device searching. In this paper we demonstrate the use of an optimized string searching algorithm to recognize exact patterns from a large database. The underlying principle in designing the algorithm is that each letter that maps to a fixed real values and some arithmetic operations which are applied to compute corresponding pattern and substring values. We have implemented this algorithm in C. We have tested the algorithm using a large dataset. We created our own dataset using DNA sequences. The experimental result shows the number of mismatch occurred in string search from a large database. Furthermore, some of the inherent weaknesses in the use of this algorithm are highlighted.