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{206935,
        author = {Narendra B O and Dr.Usha G R and Dr.Sreenivasa B R},
        title = {A Lightweight Decision Support System for Real-Time Predictive Process Monitoring},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {2},
        pages = {3130-3137},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=206935},
        abstract = {Despite the fact that predictive process monitoring (PPM) is vital for keeping tabs on normal enterprise operations, getting from a hard and fast of regulations’ theoretical correctness to placing it into workout isn't always any easy feat. no matter the truth that deep reading architectures that depend closely on computation had been the focal point of recent studies, process managers want gear which may be easy to recognize and use so that you can interfere in real-time. To fill this void, the authors of this article present an interactive DSS. The device, which is built as a surrender-to-surrender internet framework, takes raw data from event logs and makes use of dynamic trace extraction to create visualizations of usage distributions and computing case durations in real-time. a framework-based totally automated eliminate detection device is used for anomaly manipulate. This mechanism flags deviations based totally on imply-period thresholds. the use of a mild-weight dual-model structure in conjunction with prefix-based totally trace encoding, the predictive module guarantees fast inference without the want for specialised hardware. On one hand, a concurrent Linear Regression model estimates the last steps to case of completion, even as on the other, a Logistic Regression classifier is used to forecast the following probabilistic interest in an on-foot trace. With a baseline accuracy of sixty-two% and the capability to provide process stakeholders fast, actionable insights, the predictive classifier is evaluated on a multi-interest event log that includes the subsequent states: begin, examine, Approve, entire, Reject, and Revise. The counselled framework provides a without problem implementable answer to the issues of proactive workflow optimization and subsequent-satisfactory-movement recommendation via the use of placing an emphasis on computational overall performance and consumer-centric layout in preference to mathematical complexity.},
        keywords = {Predictive Process Monitoring, Process Mining, Decision Support Systems, Logistic Regression, Delay Detection, Event Log Analysis.},
        month = {July},
        }