Sipakar: An Expert System for University Major Recommendation Based on Forward Chaining Inference Engine

Sipakar is a university major recommendation expert system developed using Java Swing, based on a forward chaining inference engine and MySQL database. It is the final project for the Artificial Intelligence course at UIN Antasari Banjarmasin. The system targets high school students, university applicants, and students seeking to transfer majors. By collecting information such as personal interests, abilities, and academic background, it uses expert rules to derive personalized major recommendations, addressing the problem of students' confusion in choosing a major.

In the era of information explosion, students often feel confused when facing numerous university majors and complex career paths. As the final project of the AI course, Sipakar aims to transform the experience of education experts into an executable decision-making tool. It provides support for students' major selection through traditional expert system technology (forward chaining inference), filling the gap in the demand for intelligent tools in the education consulting field.

Forward Chaining Inference Engine

The core of the system is the forward chaining inference engine, with the following workflow:

1. Fact Collection: Obtain students' academic scores, hobbies, personality traits, career goals, etc. through questionnaires;
2. Rule Matching: Match facts with "IF-THEN" rules in the knowledge base (e.g., "IF excellent math scores AND likes logical reasoning THEN recommend science and engineering majors", "IF good at language expression AND likes communicating with others THEN recommend humanities and social science majors");
3. Forward Deduction: Gradually apply rules from known facts to derive recommendation results;
4. Explanation Mechanism: Display the inference path to enhance user trust.

Java Swing Interface

A cross-platform desktop interface is built using Java Swing, including a welcome page, multi-step questionnaire, progress indicator, result display page, and history record function.

MySQL Database

Stores the knowledge base (major information, rule sets), user data (student profiles, evaluation history), and system logs to ensure data integrity and query efficiency.

Knowledge Acquisition Sources

1. Consultation with education experts (interviews with admission teachers and career planners);
2. Academic research (literature on educational psychology and career choice theories);
3. Historical data (major selection and employment status of previous students);
4. Official materials (university major introductions, curriculum settings).

Knowledge Base Structure

Contains four types of knowledge:

• Factual knowledge: Basic major information, curriculum structure, employment prospects;
• Rule-based knowledge: Mapping relationships between conditions and recommendations;
• Heuristic knowledge: Empirical suggestions (e.g., "If math is weak but artistic talent is high, consider design-related majors");
• Meta-knowledge: Rule priorities, applicable conditions, etc. (knowledge about knowledge).

Main Application Scenarios

1. High school student college entrance guidance: Help understand their own strengths, explore suitable majors, and develop exam preparation strategies;
2. Major transfer consultation: Evaluate the matching degree between current and target majors, analyze feasibility and challenges;
3. Starting point for career planning: Introduce major career paths, industry trends, and skill development suggestions;
4. Educational institution tool: Serve as a preliminary screening tool to improve consultation efficiency, standardize evaluations to ensure quality, and collect student demand data.

Core Value

Reduce students' selection anxiety and enhance decision-making confidence; Provide efficient and standardized consultation support for educational institutions.

Current Limitations

1. Knowledge coverage: May only cover some majors of local Indonesian universities;
2. Rule completeness: Expert rules are difficult to cover all complex situations;
3. Personalization level: Lack the fine user profiling capability of modern recommendation systems;
4. Update and maintenance: Major information and employment trends need continuous updates.

Improvement Suggestions

1. Hybrid inference: Combine forward and backward chaining inference to improve flexibility;
2. Machine learning integration: Use historical data to train models to supplement rule-based inference;
3. Web-based version: Develop a web version to improve accessibility;
4. Multilingual support: Expand the user base;
5. Data visualization: Enhance the intuitiveness of result display;
6. Community function: Add user feedback mechanisms to continuously optimize recommendation quality.

Educational Significance

Sipakar embodies the importance of transforming theory into practice in AI education: through the project, students learn expert system concepts and experience the complete software development process, including requirement analysis, system design, implementation, testing, and documentation.

Project Summary

Although Sipakar uses a traditional technology stack, it successfully solves the practical problem of students choosing majors, demonstrating the unique value of rule-based inference systems in the education consulting field. Insights for AI learners: Do not have to pursue cutting-edge technologies—classic methods still have application scenarios; Domain knowledge and user needs are the keys to the success of AI applications; The initial version should be simple and leave room for expansion.

This project provides a reference architecture and implementation example for the development of similar systems.