Year: 2026
Type: preprint
Abstract: The present paper outlines the design and implementation of a symmetric substitution cipher utilizing a dynamic session key that emphasizes highly efficient and secure information processing. By incorporating character position integrity and key-conditional changes, the proposed cipher effectively resists frequency analysis and standard cryptographic attacks. Comparative evaluations demonstrate superior randomness, complexity, and computational scaling properties at larger message sizes and higher key strengths. Furthermore, Support Vector Machine (SVM) and Random Forest (RF) models are deployed to predict performance and classify cryptographic operations, highlighting the empirical strengths and limitations of modern machine learning approaches in this domain. Ultimately, the suggested cipher provides an effective alternative to established symmetric encryption algorithms, enhancing real-time execution safety within a modular cryptographic framework
Source: Research Square
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7204817262