Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{182323,
        author = {A Joshua Issac Joshua and G. Sathya and M. A. Reetha Jeyarani and B. T. Kirthika and N. Priscilla Vilma Manorathi and Mrs. Ganga Naidu and A Joshua Issac and Dr K Uthra Devi},
        title = {A Machine Learning-Driven Software Engineering Framework for Intelligent Cyber-Physical System Efficiency},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {2},
        pages = {1527-1535},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=182323},
        abstract = {Cyber-Physical Systems (CPS) represent the integration of computation, networking, and physical processes, widely adopted in critical domains such as manufacturing, healthcare, energy, and transportation. With CPS getting richer in interdependencies and more data-driven, real-time anomaly detection and the preservation of system efficiency are challenging to achieve in engineering terms. The proposed study offers a new approach to software engineering involving the use of Hybrid Inception-SVM model to be used as part of the CPS architecture to improve performance and reliability of operations. The approach covers an elaborate pipeline: collecting data on CPS sensors and actuators and preprocessing with the selection of features based on Recursive Feature Elimination (RFE), deep feature representation implemented through Inception modules, and SVM as the method of classification. Through an adaptive feedback loop the model predictions are dynamically coupled with the system control loop increasing the optimization within the system on a continual basis through real time learning. Compared with benchmark machine learning models and state-of-the-art deep learning models, the proposed Hybrid Inception-SVM model shows high performance of 97.88% accuracy, 96.78% precision, 95.67% recall, and F1- score of 96.22, within a reasonable inference time of 32.34ms. The proposed model has a more biased trade-off between accuracy and speed compared to the traditional models, like Random Forest, CNN, and LSTM, and this property renders such a model applicable in real-time applications of CPS. The integrated framework not only enhances system efficiency, but also makes CPS proactively react to dynamic environments. The results show that there is a drastic improvement of software engineering practices of CPS due to the connection with deep learning capabilities and rule-based control systems, which open the door of intelligent, self-optimizing infrastructures.},
        keywords = {Cyber-Physical Systems, Machine Learning, Inception Network, Support Vector Machine, Feature Selection, Real-Time Optimization, Efficiency Enhancement},
        month = {July},
        }