PETIR

Uji Mutasi pada Penerapan Token Mitigasi Kerentanan Cross Site Request Forgery

riva hakim — Fri, 10 Jan 2025 00:00:00 +0000

Keamanan aplikasi web merupakan perhatian kritis karena banyaknya kerentanan seperti SQL Injection (SQLi), Cross-Site Scripting (XSS), dan Cross-Site Request Forgery (CSRF). Kerentanan-kerentanan ini dieksploitasi oleh penyerang untuk mendapatkan akses yang tidak sah dan merusak aplikasi web. Penelitian kami berfokus pada analisis kerentanan Cross-Site Request Forgery (CSRF) dengan menggunakan pendekatan pengujian mutasi, yang menerapkan 5 operator mutasi yang memutasi token input pada form input. Kami memperkenalkan alat otomatis untuk mengidentifikasi dan mengatasi kerentanan CSRF menggunakan pola token rahasia. Alat ini meningkatkan keamanan aplikasi web berbasis PHP tanpa mengorbankan fungsionalitasnya. Ketika kerentanan berhasil dideteksi, aplikasi akan memberi tahu pengguna agar segera dapat diperbaiki, indikator skor mutasi kami gunakan sebagai alat pengukuran sejauh mana pengujian mutasi berhasil dilakukan, hasilnya dari 1022 mutasi yang dihasilkan seluruhnya dapat dihentikan dengan presentasi 100% menunjukan bahwa operator mutasi yang digunakan dapat bekerja dengan baik untuk mendeteksi mutan yang dihasilkan.

Penerapan YOLOv8 Dalam Deteksi Penyakit Tanaman Daun Jambu Air Secara Real-time

Nurul Safina Safina — Fri, 10 Jan 2025 00:00:00 +0000

Pendeteksian penyakit pada tanaman adalah tantangan utama dalam pertanian untuk menjaga kesehatan dan produktivitas tanaman. Penelitian ini mengimplementasikan metode YOLOv8 untuk mendeteksi penyakit pada daun jambu air dan mengevaluasi akurasi menggunakan mean Average Precision (mAP). Dataset yang terdiri dari 754 gambar daun jambu air diperluas menjadi 1229 gambar melalui proses augmentasi. Preprocessing dataset dilakukan di Roboflow, sementara pelatihan model dilakukan di google collab. Model YOLOv8 dilatih untuk mendeteksi enam jenis penyakit—embun jelaga, berlubang, antraknosa, mosaic virus, layu fusarium, dan gall—serta satu kategori daun sehat. Pengujian dilakukan terhadap 75 daun untuk memperoleh hasil mAP, yang menunjukkan nilai sebesar 82%. Sistem deteksi dibangun menggunakan framework Flask untuk integrasi dengan antarmuka pengguna berbasis web. Hasil penelitian menunjukkan bahwa YOLOv8 efektif dalam mendeteksi penyakit pada daun jambu air dengan akurasi tinggi, memberikan kontribusi signifikan dalam pengelolaan kesehatan tanaman.

Kata kunci: deteksi penyakit tanaman, YOLOv8, mean average precision, daun jambu air, framework flask

Sistem Monitoring dan Kendali Proteksi Motor 3 Phasa Berbasis IoT menggunakan Metode Fuzzy Sugeno

Lukman Medriavin Silalahi, Ridho Prananda — Fri, 10 Jan 2025 00:00:00 +0000

Masalah riset ini adalah minim pengawasan dan pengendalian motor 3 fase sebagai komponen vital dalam sistem mekanik dan kelistrikan agar setiap proses operasi menjadi aman dan efisien. Oleh sebab itu, riset ini mengusulkan pengembangan sistem motor 3 fase menggunakan metode Fuzzy Sugeno dan dikendalikan secara jarak jauh untuk mengendalikan kecepatan motor, getaran, suhu, kendali arus dan tegangan dan ditampilkan ke website. Melalui sistem ini, pengguna dapat memantau kinerja motor 3 fase secara akurat dan mengambil tindakan yang diperlukan jika terjadi gangguan atau kondisi abnormal. Sistem ini juga memungkinkan optimalisasi penggunaan energi dengan mengatur kecepatan motor sesuai kebutuhan aplikasi. Penelitian ini diharapkan dapat berkontribusi dalam meningkatkan efisiensi, keandalan, dan keamanan operasi motor 3 fase. Dengan menerapkan sistem ini, pengguna dapat mengurangi waktu henti produksi yang tidak terduga dan mengoptimalkan pemeliharaan preventif berdasarkan data yang diperoleh dari pemantauan motor secara real-time.

Sistem Rantai Pasok Buah Melon Menggunakan Teknologi Blockchain

Noer Arbian Nisyah, Irman Hermadi, Karlisa Priandana — Fri, 10 Jan 2025 00:00:00 +0000

Melon is a popular fruit and has the potential to contribute to the country's economy if its distribution and production are managed well. The supply chain requires a system that is able to present information with the characteristics of trusted, traceable and transparent. Blockchain is a technology where all shared transactions are recorded without deleting previous transaction records. This technology can assist every actor in the melon fruit supply chain in recording melon fruit transactions. Transparent transaction records can enhance consumer satisfaction and trust in the quality of the melon fruit to be consumed. This study aims to develop and implement blockchain technology in a traceability system in the melon supply chain. The method used for system development is the prototyping method. The blockchain framework used is Hyperledger Fabric. This framework was chosen because it is one of the private blockchain networks where each member can interact according to the predefined chain code. This study has effectively developed a blockchain -based prototype for a melon supply chain, known as Melon Apk. Each transaction record generates a transaction ID, which is produced in a QRcode, thereby simplifying the tracking of information and the history of the melon supply chain.

Data Driven Decision Making dalam Penentuan Kelayakan Soal Ujian Menggunakan Algoritma Fuzzy

Yudhy Setyo Purwanto, Hendra Jatnika — Fri, 10 Jan 2025 00:00:00 +0000

The need for evident-based education makes some steps that must be taken by educational institutions. In relation to improving learning and student achievement, data driven decision making (DDDM) can be utilized to find out various things, such as the level of knowledge and skills of students at a personal level, one of which is through examinations. This research was conducted to determine the eligibility standards of college-level English exam questions using DDDM and fuzzy algorithms. 20 first-year students were tested to see the quality of the questions based on the results of the exam. The results are the level of difficulty of the question 0.611, Discrimination Power 20.2%, Validity produces 19 valid questions and 31 invalid questions, and reliability results are reliable. The conclusion is that DDDM method can be used to calculate the feasibility level of questions so that it can provide input on the development of exam questions standardization, especially for English. Further research should be conducted with larger data in several locations with different participants, thus the standard of questions’ eligibility can be more accurate. The use of different algorithms or calculation models is also recommended so that it can provide benefits to this field.

Pemodelan Fuzzy Inference System (FIS) dan Certainty Factor (CF) untuk Grading Ternak pada Penggemukan Sapi Bali

Defiana Arnaldy, Kudang B. Seminar, Muladno, Heru Sukoco, Shelvie Nidya Neyman — Fri, 10 Jan 2025 00:00:00 +0000

Beef cattle development in Indonesia has developed in several regions by applying livestock technology and innovation through the school for smallholder communities (namely SPR in Indonesian). The SPR program has been established in several regions in Indonesia and has had a positive impact on smallholder farming. The condition of livestock in each SPR is still classified as not having good productivity. One of the most important factors in developing livestock populations for breeding and fattening purposes is the availability of good seed sources based on good genetic quality. The grading classification of cattle for breeding and fattening purposes needs to be identified in detail and comprehensively, including through selection based on qualitative and quantitative traits. The main objective of this research is to develop a livestock grading model in Balinese cattle fattening using fuzzy inference system and certainty factor approach. Grading is done by looking at information from the morphometric characteristics of livestock. The method used in this research is quantitative and qualitative data collection with a system of direct interviews with farmers and measurements of their livestock. The parameters used follow the characteristics of animal morphometrics, namely body weight, body length, chest circumference, chest width, hip height, hip width, pelvic height, hip height. The results obtained from the initial data of the experiment show that the classification is divided into three classes, namely class I, class II, and class III.

Mitigasi Pembangkitan Token Pada Kerentanan Cross Site Request Forgery dengan Pengujian Mutasi

Fri, 10 Jan 2025 00:00:00 +0000

In the digital era, software security is crucial to prevent financial losses and data theft due to cyber attacks such as Cross-Site Request Forgery (CSRF). Based on CVEdetails data, CSRF attacks increased significantly from 416 cases in 2020 to 1398 cases in 2023. This research explores the use of Static Application Security Testing (SAST) to detect and prevent CSRF attacks. In addition, anti-CSRF token and mutation testing are used to improve the quality of test cases in detecting CSRF vulnerabilities. The research results show that the mutation testing technique is effective in improving the quality of test cases, with the Mutation Score Index (MSI) value increasing from 50% to 100% after improvement. This research concludes that mutation testing can improve the quality of test cases, thereby providing better software protection against CSRF attacks.

Optimalisasi Prediksi Afinitas Interaksi Obat-Target dengan Graph Neural Network dan Attention Mechanism

Husni Fadhilah Dhiya Ul Haq, Pawesi Siantika, Dimmas Mulya, Putri Saptawati — Fri, 10 Jan 2025 00:00:00 +0000

Virtual screening pada obat memainkan peran penting dalam meningkatkan throughput penemuan dan mengurangi biaya R&D. Deep learning muncul sebagai solusi menjanjikan, menawarkan hasil yang efisien tanpa memerlukan keahlian domain yang luas atau detail struktural. Studi ini memperkenalkan iGanDTA, sebuah perbaikan dari model multitask yang mampu memprediksi afinitas pengikatan obat-target dengan akurasi tinggi dan mengklasifikasikan interaksi dengan performa yang sangat baik. Menggunakan residual graph neural network, iGanDTA memproses data fingerprint dari senyawa untuk membedakan tingkat pengikatan dalam urutan protein. Evaluasi pada dataset benchmark menunjukkan performa yang superior dibandingkan dengan metode yang ada, dengan iGanDTA mencapai skor MSE, CIndex, dan R2 masing-masing 0.238, 0.894, dan 0.710 pada dataset Davis serta 0.181, 0.864, dan 0.746 pada dataset KIBA

Pendekatan Metode Forward Chaining Pada Penentuan Menu Makanan Sehat Pasien Paru Obstruktif Kronis

Abdurrasyid, Meilia Nur Indah Susanti, Indrianto, Syahrun Mubarak — Fri, 10 Jan 2025 00:00:00 +0000

Penyakit kronis biasanya disebabkan akibat pola hidup yang kurang sehat, seperti merokok, makan makanan berlemak, makanan cepat saji, cara diet yang tidak tepat, serta kurang berolahraga. Salah satu dari tiga penyakit yang menyebabkan kematian terbesar di dunia adalah penyakit paru-paru kronis yang dikatakan tidak dapat disembuhkan. Penyakit ini menjadi masalah bagi sistem kesehatan dunia. Penyakit PPOK muncul dengan ditandai gejala pernapasan jangka panjang dan keterbatasan aliran udara yang tidak dapat pulih dengan baik dan biasanya dapat menjadi lebih parah. Merokok adalah penyebab utama penyakit tersebut. Petugas gizi di Rumah Sakit Umum Kesrem Lhokseumawe saat ini menghadapi kesulitan dalam menyusun menu makanan pasien karena mereka masih menggunakan metode manual, meskipun mereka telah membaca buku pedoman pelayanan gizi rumah sakit. Akibatnya, menentukan menu makanan yang akan diberikan kepada pasien menjadi sulit dan memakan waktu. Oleh karena itu, diperlukan adanya sistem pengambilan keputusan makanan pada pasien PPOK berbasis web yang menggunakan sistem pakar dengan menggunakan pendekatan forward chaining. Metode ini digunakan untuk memberikan rekomendasi untuk penentuan makanan bergizi. Ketika menerapkan metode ini dalam memberikan rekomendasi penentuan makanan bergizi, diperoleh hasil yang memuaskan setelah 20 kali pengujian, dan akurasinya mencapai 100% sesuai kriteria pasien. Penerapan metode forward chaining, diharapkan petugas gizi dapat lebih mudah menyusun menu makanan yang sehat.

Implementasi Internet of Things dan Deteksi Anomali Menggunakan Algoritma Deep Learning Pada Distribusi Buah Melon

Andi Sulviqrah Amalia, Karlisa Priandana, Irman Hermadi — Fri, 10 Jan 2025 00:00:00 +0000

Melon is the horticultural plants that had the potential to improve Indonesia's economy in the agricultural sector. Efforts to improve the economy must been accompanied by improvements in the quality of melon fruit production and distribution to reach consumers. One way to maintain the quality of melon fruit is to combine production and distribution processes with the use of temperature sensors. Utilizing temperature sensor with Internet of Things (IoT) technology to monitor melon temperatures during the distribution process is a form of technological innovation. This research aims to develop a melon distribution system by applied IoT devices to monitor environmental temperature and detect anomalies before transferring data to blockchain system. The anomaly detection method in this research uses deep learning algorithms. Autoencoder was chosen as the architecture model in this research because this method can help minimize data anomalies. The results of this research indicate that IoT technology and anomaly detection were successfully implemented and performed very well. Based on performance testing using quality of service parameters, the throughput was 86,152 bps, the delay was 0.041199 ms, and the packet loss was 0.064%. The evaluation results of anomaly detection model for precision, recall, and F1-score were 0.9952, 1, and 0.9658.

Knowledge Management System Balai Pelestarian Kebudayaan Menggunakan Metode KMSLC

Fri, 10 Jan 2025 00:00:00 +0000

ABSTRACT

In a dynamic and competitive era, knowledge is considered a crucial element for the survival of an organization. The ability to effectively manage knowledge is a determining factor for organizational success. Therefore, the application of knowledge management is crucial for competitive advantage in the information age. The application of Knowledge Management also penetrates into various types of organizations, including government agencies which are driven by factors such as employee movement or retirement. The Balai Pelestarian Kebudayaan Wilayah V faces the problem of limited Cultural Pamong due to the high demand from related agencies both at the central and regional government levels. Without a knowledge management system in the Balai Pelestarian Kebudayaan Wilayah V, the knowledge and experience of cultural staff will be lost along with the movement of cultural staff. The development of a Knowledge Management System using the KMSLC method aims to become a virtual space for storing and sharing knowledge among cultural leaders. Functionality testing with blackbox testing shows that the features in the KMS have run well. Then usability testing using SUS from KMS obtained acceptable results in the acceptability ranges indicator and got a good predicate in adjective ratings.

Komparasi Algoritme Machine Learning Untuk Prediksi Pemeliharaan Preventif Precision Air Conditioning di Data Center

Kahfi Suradiradja, Dani Ramdani, Karno Nano — Fri, 10 Jan 2025 00:00:00 +0000

Data Center atau pusat data merupakan fasilitas dasar untuk penerapan suatu layanan teknologi informasi yang menerapkan tren teknologi terkini harus terjamin ketersediaan layanannya selalu dapat diakses, sehingga dibutuhkan solusi pendukung sebagai bagian dari strategi pemeliharaan perangkat dan proses pemantauan operasional dalam rangka mempertahankan ketersediaan layanan data center. Pada penelitian ini membandingkan beberapa algoritme klasifikasi machine learning untuk memprediksi kondisi perangkat Precision Air Conditioning pada operasional data center. Dataset untuk penelitian ini adalah data log harian Precision Air Conditioning di Badan Riset dan Inovasi Nasional (BRIN). Tujuan dari penelitian ini, mengidentifikasi beberapa algoritme klasifikasi machine learning yakni Decision Tree, Random Forest, Artifcial Neural Network Multi Layer Perceptron, Naïve Bayes dan Support Vector Machine. Tahapan dari metode penelitian ini yakni analisis pemahaman terhadap masalah pengklasifikasian status pemeliharaan, pengambilan data data log Precision Air Conditioning, kemudian dilanjutkan proses beberapa pemodelan algoritme – algoritme machine learning dan evaluasi untuk mendapatkan model algoritme dengan hasil akurasi yang baik. Hasil pengukuran pada evaluasi dari beberapa machine learning pada penelitian ini menghasilkan model decision tree C4.5 memiliki tingkat akurasi terbaik yakni 98,75 persen. Role yang dihasilkan dapat digunakan untuk memprediksi kondisi perangkat di data center sebagai tindakan preventif.

Performance Comparison of VGG16, Mobilenet, And Xception Model Architecture in Rice Plant Leaf Identification

Riki Ruli A. Siregar, Nurin Masyitha Safiera, Abdul Haris — Fri, 10 Jan 2025 00:00:00 +0000

Rice is one of the world's most important staple foods. Rice is a staple food in most regions of the world, especially in Indonesia. Rice plant nutrition is one of the most important things in plant growth and development. Nutrient deficiencies in plants can affect the growth process and the quality of the plants when they are ready to be harvested. The dataset used in this research comes from the Kaggle platform, which has a total of 1190 datasets. The rice leaf images are divided into 2 classes, namely Sufficient and Deficient, which are tested with a ratio of 80% training data and 10% test data, and 10% as validation data. The model architecture used in this research is 3 VGG16, MobileNet and Xception using Jupyter and Google Collaboratory as tools. The tests were performed using 10 epochs and batch sizes of 32 and 64. The best accuracy results obtained are 78.15% and 76.47% for VGG16, 82.69% and 86.55% for MobileNet, 82.33% and 88.24% for Xception. Meanwhile, the best overall accuracy result was achieved by the Xception model at 88.24% with an input batch size of 32 and the tool used was Jupyter.