<![CDATA[Gangguan layanan IndiHome sering terjadi dan memerlukan penanganan cepat. Namun, keterbatasan akses langsung ke teknisi menyebabkan pelanggan mengalami keterlambatan dalam mendapatkan solusi. Sistem pakar berbasis Case-Based Reasoning (CBR) dapat menjadi solusi untuk membantu pelanggan mengidentifikasi jenis gangguan secara mandiri. Penelitian ini bertujuan membangun sistem pakar diagnosis gangguan IndiHome menggunakan metode CBR yang mampu merekomendasikan solusi berdasarkan kemiripan kasus sebelumnya. Sistem dirancang menggunakan pendekatan CBR dengan empat tahapan utama: Retrieve, Reuse, Revise, dan Retain. Penghitungan kemiripan (similarity) dilakukan menggunakan metrik Jaccard Coefficient, dengan bobot atribut berdasarkan kepentingan gejala. Sistem diimplementasikan berbasis web menggunakan PHP dan MySQL. Validasi dilakukan melalui pengujian terhadap 50 kasus gangguan nyata dari Witel Cirebon. Hasil pengujian menunjukkan sistem mampu mengidentifikasi jenis gangguan dengan akurasi rata-rata 86%. Kasus dengan nilai similarity tertinggi digunakan sebagai dasar rekomendasi solusi, seperti pemeriksaan kabel fiber, restart modem, atau kontak ke 147. Sistem pakar berbasis CBR terbukti efektif sebagai alat bantu diagnosis awal gangguan IndiHome, memberikan solusi cepat dan akurat bagi pelanggan, serta mengurangi beban layanan pelanggan. Abstract IndiHome service disruptions frequently occur and require prompt handling. However, limited direct access to technicians causes customers to experience delays in obtaining solutions. A Case-Based Reasoning (CBR) expert system can be a solution to help customers independently identify the type of disruption. This study aims to build an expert system for diagnosing IndiHome disruptions using the CBR method that is able to recommend solutions based on the similarity of previous cases. The system is designed using the CBR approach with four main stages: Retrieve, Reuse, Revise, and Retain. Similarity calculations are performed using the Jaccard Coefficient metric, with attribute weights based on the importance of symptoms. The system is implemented web-based using PHP and MySQL. Validation was carried out through testing on 50 real disruption cases from Witel Cirebon. The test results showed the system was able to identify the type of disruption with an average accuracy of 86%. Cases with the highest similarity value were used as the basis for solution recommendations, such as checking the fiber cable, restarting the modem, or contacting 147. The CBR-based expert system has proven effective as an early diagnosis tool for IndiHome disruptions, providing fast and accurate solutions for customers, and reducing the burden on customer service..]]>
