Detecting the penetration of malicious behavior in big data using hybrid algorithms

Information security must be maintained because the amount of data in the world today is growing exponentially. The issues related to security are growing as big data usage increases. Finding ways to identify intrusions into networks and information systems is one of the major issues in this subject...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Signal, image and video processing Ročník 18; číslo Suppl 1; s. 919 - 933
Hlavní autoři: Wang, Yue, Shi, Yan
Médium: Journal Article
Jazyk:angličtina
Vydáno: London Springer London 01.08.2024
Springer Nature B.V
Témata:
Accuracy
Algorithms
Big Data
Computer Imaging
Computer networks
Computer Science
Cybersecurity
Decision trees
Efficiency
False alarms
Image Processing and Computer Vision
Immune system
Information systems
Intrusion detection systems
Machine learning
Multimedia Information Systems
Neural networks
Optical density
Optimization
Original Paper
Pattern Recognition and Graphics
Signal,Image and Speech Processing
Support vector machines
Vision
Big data
Malicious behavior
Support vector machine algorithm
Intrusion detection
Bacterial system algorithm
ISSN:1863-1703, 1863-1711
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Information security must be maintained because the amount of data in the world today is growing exponentially. The issues related to security are growing as big data usage increases. Finding ways to identify intrusions into networks and information systems is one of the major issues in this subject. It is imperative and important to enhance intrusion detection skills in order to address malevolent behavior in large data. This paper presents a scalable approach to harmful data detection. Three variables have been considered in this strategy and model: scalability, user review, and temporal progress. High volumes of data can be processed using this technology. Time is split into time periods for data training in this system, and each time interval uses users’ review information to train the data. Large volumes of data require sophisticated strategies to handle, and scalability in storage allows for faster processing and fewer computations. This approach is a kind of hardware–software hybrid solution for malware detection. A fresh approach to feature extraction has also been applied. In the suggested method, the bacteria algorithm in conjunction with the immune system algorithm has been utilized for the prediction operation, and the modified support vector machine algorithm and optical density have been utilized for classification. Based on the findings, the suggested combination algorithm outperforms other comparable techniques with a 21% detection rate, a 62% false alarm rate, a 15% accuracy rate, and a 73% training duration.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1863-1703
1863-1711
DOI:10.1007/s11760-024-03203-3