Dynamics User Identification Using Pairwise Client Coupling
Journal Title: Bonfring International Journal of Software Engineering and Soft Computing - Year 2019, Vol 9, Issue 2
Abstract
In this paper due to the increasing vulnerabilities in the internet, security alone isnt sufficient to keep a rupture, however digital crime scene investigation and cyber intelligence is also required to prevent future assaults or to identify the potential attacker. The unpretentious and incognito nature of biometric information gathering of keystroke dynamics has a high potential for use in cyber forensics or cyber intelligence and crime scene investigation or digital knowledge. The keystroke dynamics is a biometric assumption that different people typify in a unique way. The information accessing from computer systems is normally controlled by client accounts with usernames and passwords. If the set of data falls into the wrong hands, such a scheme has little security. For example fingerprints, can be used to strengthen security, however they require very expensive additional hardware. Keystroke dynamics with no additional hardware can be used. Keystroke dynamics is for the most part applicable to verification and identification also possible. In verification it is known who the client is supposed to be and the biometric system should verify if the user is who he claims to be in identification, the biometric The system should identify the client with keystroke dynamics without additional knowledge. This paper examines the usefulness of keystroke dynamics to determine the users identity. We propose three plans for user identification when entering a keyboard. We use different machine learning algorithms in conjunction with the proposed user coupling technology. In particular, we show that combined user coupling in a bottom - up tree structure scheme provides the best performance in terms of both precision and time complexity. The techniques proposed are validated by keystroke data. Lastly, we also examined the performance of the identification system and demonstrated that the performance was not optimal, as expected.
Authors and Affiliations
Tejaswi D, Rajasekar Rangasamy
Keywords
Related Articles
- EP ID EP588025
- DOI 10.9756/BIJSESC.9023
- Views 107
- Downloads 0
Heart Disease Prediction System Using Supervised Learning Classifier
Cardiovascular disease remains the biggest cause of deaths worldwide and the Heart Disease Prediction at the early stage is importance. In this paper Supervised Learning Algorithm is adopted for heart disease prediction...
Hybridization Approach to Classify Big Data Using Social Internet of Things
Recently there has been quite a number of independent research activities that investigated the potentialities of integrating social networking concepts into Internet of Things solutions. The resulting paradigm, named So...
Enhanced Secure Big Data in Distributed Mobile Cloud Computing Using Fuzzy Encryption Model
This paper proposes a novel architecture for adaptive encryption of public cloud databases that offers a proxy-free alternative to the system. cloud system is difficult to synchronize login and authentication data betwee...
Review of Data Mining Classification Techniques
A Classification is one the most valuable critical techniques. In huge information Classification procedures is helpful to deal with substantial measure of information. Orders are utilized to foresee straight out class n...