Claim Missing Document
Check
Articles

Found 3 Documents
Search

RUBBER PLANT DISEASE DIAGNOSIS SYSTEM USING DEPTH FIRST SEARCH AND CERTAINTY FACTOR METHOD Zunita Wulansari; Mukh Taofik Chulkamdi
JOSAR (Journal of Students Academic Research) Vol 6 No 1: March 2021
Publisher : Universitas Islam Balitar

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35457/josar.v6i1.1446

Abstract

Rubber plants have a very important role in the economy in Indonesia, because many people depend on this commodity. The area of ​​rubber plantations in Indonesia has reached more than 3 million hectares, while Malaysia and Thailand, which are Indonesia's main competitors, have a rubber plantation area below that number. Only 15% of the rubber area is large plantations, while 85% is smallholder plantations which are managed simply as is, some even rely on natural growth. The problems faced by rubber farmers are disease and treatment problems. With these conditions, the researcher aims to build an expert system application for the diagnosis of rubber plant diseases by applying the depth first search method and Certainty Factor is used so that the expert system can reason like an expert, and to get the highest confidence value. The problems faced by rubber farmers are disease and treatment problems. Given these conditions, the researcher aims to build an expert system application for the diagnosis of rubber plant diseases by applying the depth first search certainty factor method. Depth first search and Certainty Factor methods are used so that the expert system can reason like an expert, and to get the highest confidence value. The application design by applying the depth first search method and certainty factor was successfully built into a web-based application. Black box testing on this application system has been successful in accordance with the design that has been made. The test results by experts on the identification system are in accordance with direct identification. And the results of beta testing produce a percentage of 83%, which means that users have a high level of satisfaction with the application. With the results of this test, the selected diseases were fungus disease with an accuracy of 87.54%, spot cancer with an accuracy of 97.64% and root rot disease with an accuracy of 97.41%.
First Aid Diagnosis Expert System Using the Certainty Factor Method Fredy Dwi Aditia; Mukh Taofik Chulkamdi
JOSAR (Journal of Students Academic Research) Vol 6 No 1: March 2021
Publisher : Universitas Islam Balitar

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35457/josar.v6i1.1463

Abstract

Accidents often occur suddenly and are not planned, which can result in injury. If an accident occurs, first aid must be carried out immediately before being given further assistance. The Indonesian Red Cross (PMI) is one of the institutions in Indonesia engaged in the health sector that can provide first aid in the event of an accident. However, often when an accident occurs, the victim is not immediately given help due to lack of information to the public. In this study, the authors made an application of the Certainty Factor method so that the public could recognize and know what to do with victims who had an injury or accident. This application is made to present the knowledge of an expert in approaching a problem, which is called an expert system. This expert system will display symptoms that can be selected according to the symptoms felt by the victim. The final results of this application obtained a value of 78% from the calculation using the Certainty factor method in bleeding diseases. From the tests carried out, it shows the beta test results obtained by a value of 58.25% indicating that the respondents strongly agree with the usefulness of the application system
Comparative Study Of Monthly Electricity Consumption Clusterization Using K-Means and DBSCAN Putri Rochfiani; Mukh Taofik Chulkamdi; Udkhiati Mawaddah
JOSAR (Journal of Students Academic Research) Vol 10 No 2 (2025): September
Publisher : Universitas Islam Balitar

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35457/rqvgbf18

Abstract

Menganalisis pola konsumsi listrik sangat penting untuk meningkatkan efisiensi distribusi energi dan mengidentifikasi anomali seperti lonjakan yang tidak biasa atau kemungkinan pencurian listrik. Penelitian ini menyajikan analisis komparatif dua algoritma klaster—K-Means dan DBSCAN—dalam mengklasifikasikan penggunaan listrik bulanan pelanggan PT PLN (Persero) Rayon Ngunut, yang mencakup Kecamatan Rejotangan, Ngunut, Kalidawir, dan Pucanglaban. Dataset tersebut mencakup catatan konsumsi dari bulan November dan Desember 2024. Algoritma K-Means, yang menggunakan pendekatan klaster berbasis centroid, bekerja efektif pada dataset yang seragam, sementara DBSCAN, sebuah metode berbasis kepadatan, lebih mampu mengenali outlier dan pembentukan klaster yang tidak sferis. Kinerja kedua algoritma dievaluasi menggunakan Akurasi, Mean Squared Error (MSE), Presisi, Recall, dan F1-Score. Hasil eksperimen menunjukkan bahwa K-Means mencapai akurasi 96%, MSE 0,0400, presisi 0,71, recall 1,00, dan skor F1 0,83. Sebaliknya, DBSCAN mencapai akurasi 76%, MSE 0,2400, presisi 0,29, recall 1,00, dan skor F1 0,45. Hasil ini menunjukkan bahwa K-Means menghasilkan klaster yang lebih kompak dan konsisten, sementara DBSCAN lebih unggul dalam mengidentifikasi anomali, dengan total mendeteksi 17 outlier. Akibatnya, K-Means dianggap lebih cocok untuk pengelompokan konsumsi yang stabil, sedangkan DBSCAN direkomendasikan untuk tujuan deteksi anomali. Temuan ini diharapkan dapat membantu PT PLN (Persero) dalam mengembangkan strategi berbasis data dan adaptif untuk manajemen energi yang lebih efisien.