Perancangan Sistem Inferensi dan Pengambil Keputusan Berbasis Aturan untuk Rekomendasi Jalur Pembelajaran pada Platform Founderhub

Heriswan, Isaura Qinthara (2026) Perancangan Sistem Inferensi dan Pengambil Keputusan Berbasis Aturan untuk Rekomendasi Jalur Pembelajaran pada Platform Founderhub. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5026221146-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Pertumbuhan platform pembelajaran daring telah memperluas akses terhadap materi kewirausahaan, namun juga menimbulkan permasalahan information overload dan kebingungan pengguna dalam menentukan jalur pembelajaran yang sesuai, khususnya bagi founder baru yang belum memiliki data historis aktivitas. Kondisi ini dikenal sebagai permasalahan cold start pada sistem rekomendasi. FounderHub sebagai platform pembelajaran kewirausahaan yang mengadopsi model peran founder 4H (Hustler, Hipster, Hacker, dan Handler) memerlukan mekanisme rekomendasi yang mampu memberikan jalur pembelajaran yang relevan, terstruktur, dan dapat dijelaskan sejak awal pengguna bergabung. Tugas Akhir ini bertujuan merancang dan mengimplementasikan sistem pengambil keputusan berbasis sistem pakar untuk menghasilkan rekomendasi jalur pembelajaran personal pada platform FounderHub. Sistem dibangun menggunakan metode penalaran Forward Chaining yang memproses hasil asesmen 4H dan Learner Persona sebagai fakta awal menuju rekomendasi melalui basis pengetahuan yang terdiri atas 39 aturan IF–THEN dalam empat layer inferensi. Arsitektur sistem mengintegrasikan Assessment Engine, Working Memory, Rule Base, Curriculum Knowledge Base, Inference Engine, Recommendation Engine, dan Explainability Layer sehingga mampu menghasilkan rekomendasi yang bersifat deterministik, transparan, dan mendukung prinsip Explainable Artificial Intelligence (XAI). Metode yang digunakan adalah system design research yang meliputi studi literatur, pengumpulan dan analisis data, analisis kebutuhan, perancangan basis pengetahuan, perancangan mekanisme penalaran, implementasi model inferensi, serta evaluasi sistem melalui unit testing, black-box testing, validasi pakar, dan umpan balik pengguna. Hasil implementasi menunjukkan bahwa seluruh komponen sistem berhasil dijalankan sesuai rancangan. Pengujian black-box terhadap 14 skenario memperoleh tingkat keberhasilan 100%, seluruh 63 unit test berhasil dijalankan tanpa kegagalan, serta batch testing terhadap 120 data asesmen berhasil diproses tanpa kesalahan maupun aktivasi aturan fallback. Validasi oleh dua orang pakar memperoleh nilai rata-rata 4,41 dari 5,00, sedangkan evaluasi terhadap lima pengguna memperoleh nilai rata-rata 4,45 dari 5,00 yang menunjukkan bahwa sistem dinilai sangat baik dalam menghasilkan rekomendasi yang relevan, konsisten, dan mudah dipahami.

================================================================================================================================

The rapid growth of online learning platforms has expanded access to entrepreneurship education but has also introduced challenges such as information overload and user uncertainty in selecting appropriate learning paths, particularly for new founders who lack historical interaction data. This condition is commonly referred to as the cold start problem in recommender systems. FounderHub, an entrepreneurship learning platform adopting the 4H founder role model (Hustler, Hipster, Hacker, and Handler), requires a recommendation mechanism capable of providing relevant, structured, and explainable learning paths from the user's first interaction. This undergraduate thesis aims to design and implement a rule-based expert system that generates personalized learning path recommendations for the FounderHub platform. The proposed system employs the Forward Chaining reasoning method to process users' 4H assessment results and Learner Persona as initial facts through a knowledge base consisting of 39 IF–THEN rules organized into four inference layers. The system architecture integrates an Assessment Engine, Working Memory, Rule Base, Curriculum Knowledge Base, Inference Engine, Recommendation Engine, and Explainability Layer, enabling deterministic, transparent, and explainable recommendations that align with the principles of Explainable Artificial Intelligence (XAI). The study adopts a system design research methodology consisting of literature review, data collection and analysis, requirements analysis, knowledge base design, inference mechanism design, inference model implementation, and system evaluation through unit testing, black-box testing, expert validation, and user feedback. The implementation results demonstrate that all system components functioned as designed. Black-box testing across 14 scenarios achieved a 100% success rate, all 63 unit tests passed successfully, and batch testing on 120 assessment records completed without errors or fallback rule activation. Expert validation involving two experts achieved an average score of 4.41 out of 5.00, while user evaluation involving five participants obtained an average score of 4.45 out of 5.00, indicating that the proposed system provides relevant, consistent, and easily understandable recommendations.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Knowledge-Based Recommender System, Forward Chaining, Sistem Pakar Berbasis Aturan, Jalur Pembelajaran Personal, Explainable Artificial Intelligence, Rule-Based Expert System, Personalized Learning Path.
Subjects:	T Technology > T Technology (General) T Technology > T Technology (General) > T58.62 Decision support systems
Divisions:	Faculty of Information Technology > Information System > 57201-(S1) Undergraduate Thesis
Depositing User:	Isaura Qinthara Heriswan
Date Deposited:	01 Aug 2026 01:57
Last Modified:	01 Aug 2026 01:57
URI:	http://repository.its.ac.id/id/eprint/141249

Actions (login required)

View Item