High Impact Factor : 4.396 icon | Submit Manuscript Online icon |

An Intelligent Web-Based Adaptive Examination Framework with Inline Cognitive Stress Detection Using Native JVM Machine Learning

Author(s):

Pratiksha Gadekar , PES Modern College of Engineering, Pune; Vaishnav Mohare, PES Modern College of Engineering, Pune; Vaishnavi Fasale, PES Modern College of Engineering, Pune; Chandan Vaishnav, PES Modern College of Engineering, Pune; Mr. Shripad Bhide, PES Modern College of Engineering, Pune

Keywords:

Computer Adaptive Testing, Item Response Theory, Behavioural Biometrics, Cognitive Stress Index, Native JVM Machine Learning, Spring Boot, Smile, Random Forest

Abstract

Conventional online examinations present every candidate with the same static question sequence and judge performance by the final score alone, ignoring the cognitive strain a candidate experiences during the test. Existing stress-detection approaches mostly depend on cameras or wearable sensors, and behaviour-based systems usually hand the captured telemetry to a separate service written in another language, which adds network latency and exposes sensitive data in transit. This paper presents a web-based adaptive examination framework that combines Computer Adaptive Testing based on Item Response Theory with the estimation of a Cognitive Stress Index from browser-side behavioural telemetry (key-hold time, key-flight time, cursor-hover duration, option toggling and navigation hesitation), without using a camera or microphone. A Random Forest, trained offline and exported as decision-tree weights, is held in the memory of the Java Virtual Machine (JVM) backend so that inference needs no network hop. A prototype with a student examination interface and an administrative telemetry dashboard is described. In the reported prototype tests, in-process inference took well under a millisecond and the stress classifier showed high agreement with its labels on a generated dataset; end-to-end concurrent-user tests and validation on real examinees remain future work.

Other Details

Paper ID: IJSRDV14I90006
Published in: Volume : 14, Issue : 9
Publication Date: 01/12/2026
Page(s): 6-12

Article Preview

Download Article