A Distinguished Method for Network Intrusion Detection using Random Initialized Viterbi Algorithm in Hidden Markov Model

Intrusion Detection System (IDS) is a system that surveils the dubious network activity. There are several approaches which deal with intrusion detection and cyber-attack detection, but the most optimal IDS would be the one which can predict the upcoming threats along with detecting the present atta...

Full description

Saved in:
Bibliographic Details
Published in:2022 OITS International Conference on Information Technology (OCIT) pp. 273 - 277
Main Authors: P, Sharmila S, Shukla, Pratyush, Chaudhari, Narendra S
Format: Conference Proceeding
Language:English
Published: IEEE 01.12.2022
Subjects:
Computational modeling
Hidden Markov Model
Hidden Markov models
Intrusion detection system
Machine learning
Network attack prediction
Network intrusion detection
pattern recognition
Predictive models
Speech analysis
Viterbi algorithm
Viterbi decoding algorithm
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Intrusion Detection System (IDS) is a system that surveils the dubious network activity. There are several approaches which deal with intrusion detection and cyber-attack detection, but the most optimal IDS would be the one which can predict the upcoming threats along with detecting the present attacks. The Machine Learning probabilistic models work remarkably in prediction of threats among these models, Hidden Markov Model (HMM) outperforms all other models. HMM is widely used in cryptanalysis, gene prediction, computational linguistic, speech analysis as well as its synthesis and network attacks detection and prediction. In this paper, we have proposed a distinct methodology using Viterbi algorithm of HMM which is initialized with the random parameter. Our methodology significantly upsurges the detection accuracy of the current state along with all states, it also enhances the prediction accuracy of the next feasible state when compared to existing approaches.
DOI:10.1109/OCIT56763.2022.00059