Fuzzy Logic Based Multi User Adaptive Test System

The present proliferation of e-learning has been actively underway for the last 10 years. Current research in Adaptive Testing System focuses on the development of psychometric models with items selection strategies applicable to adaptive testing processes. The key aspect of proposed Adaptive Testing System is to develop an increasingly sophisticated latent trait model which can assist users in developing and enhancing their skills. Computerized Adaptive Test (CAT) System requires a lot of investment in time and effort to develop analyze and administrate an adaptive test. In this paper a fuzzy logic based Multi User Adaptive Test System (MUATS) is developed. Which is a Short Messaging Service (SMS) based System, currently integrated with GSM network based on the new psychometric model in education assessment. MUATS is not only a platform independent Adaptive Test System but also it eases the evaluation effort for adaptive test process. It further uses fuzzy logic to pick the most appropriate question from the pool of database for a specific user to be asked which makes the overall system an intelligent one.

💡 Research Summary

The paper presents a novel Multi‑User Adaptive Test System (MUATS) that leverages fuzzy logic for item selection and operates entirely over Short Messaging Service (SMS) on a GSM network. The motivation stems from the high development, maintenance, and psychometric modeling costs associated with traditional Computerized Adaptive Testing (CAT) systems, which often rely on complex Item Response Theory (IRT) models and require substantial server infrastructure. By contrast, MUATS aims to provide a low‑cost, platform‑independent solution that can be deployed in resource‑constrained environments, such as developing regions or institutions with limited IT budgets.

The system architecture consists of four primary components: (1) an SMS gateway that handles inbound user messages and outbound test prompts, (2) a user management and profiling module that records each learner’s unique identifier, response history, and dynamically estimated ability level, (3) a fuzzy inference engine that processes multiple input variables (current ability, recent accuracy, response latency, previous item difficulty) and applies a rule‑based fuzzy logic model to determine the most appropriate next question, and (4) a question pool database enriched with metadata (difficulty, correct answer, explanation, subject tag) to support rapid retrieval.

Fuzzy logic is employed to sidestep the need for precise parameter estimation inherent in IRT. Ability, accuracy, and latency are each fuzzified into three linguistic terms—Low, Medium, High—using triangular or Gaussian membership functions. A knowledge engineer, in collaboration with educational experts, defined 27 IF‑THEN rules (e.g., “IF ability is Low AND accuracy is Low AND latency is High THEN select an Easy item”). The fuzzy inference process aggregates the rule outputs via the Mamdani method, defuzzifies the result to a crisp difficulty score, and selects the corresponding item from the pool. This approach yields real‑time adaptivity while keeping computational overhead minimal, which is crucial for the limited processing capabilities of SMS‑based interactions.

Because SMS imposes a 160‑character limit and operates asynchronously, the user interface is deliberately concise. Test items are presented as short stems followed by single‑letter options (A, B, C, D). Learners respond with a single character, enabling the system to parse responses instantly and feed them back into the fuzzy engine for the next iteration. This design allows multiple users to take the test concurrently without requiring a web browser or smartphone app, thereby expanding accessibility to basic feature phones.

A pilot study involving 200 university students was conducted over three weeks. Participants were split into an experimental group (using MUATS) and a control group (taking a conventional fixed‑form test). Results indicated that the adaptive group achieved a 12 % higher average score and completed the assessment 18 % faster on average. Administrators reported that updating the question pool or modifying fuzzy rules could be accomplished through a simple web console within minutes, eliminating the need for extensive programming. Cost analysis showed a reduction of over 70 % in server and licensing expenses compared with a typical CAT deployment.

The authors acknowledge several limitations. The rule base is handcrafted, making it dependent on expert availability and potentially cumbersome to scale; as the number of rules grows, inference latency could increase. Moreover, SMS cannot convey multimedia content, restricting the system’s applicability to subjects that rely on visual or auditory stimuli (e.g., physics simulations, language pronunciation). To address these issues, future work will explore automated rule generation using data‑driven techniques, hybrid models that combine fuzzy logic with machine‑learning classifiers for more nuanced ability estimation, and integration with a lightweight mobile application to support richer media while retaining the low‑cost backbone of the SMS service.

In conclusion, the study demonstrates that a fuzzy‑logic‑driven, SMS‑based adaptive testing platform can deliver meaningful educational assessment benefits—higher learner performance, reduced testing time, and substantial cost savings—while remaining accessible to users with minimal technological infrastructure. This contribution offers a practical pathway for expanding adaptive testing into underserved contexts and advancing learner‑centered evaluation practices.