Pseudocode

Pendekatan Ensemble Learning untuk klasifikasi serangan DDoS

Muhammad Oriza Nurfajri, Guntur Budi Herwanto — Sat, 19 Jul 2025 00:00:00 +0000

This research proposes an ensemble learning approach for classifying Distributed Denial of Service (DDoS) attacks using the CIC-DDoS2019 dataset. DDoS attacks remain a significant threat to network security, necessitating efficient detection methods. We developed an ensemble model combining Random Forest, Gradient Boosting, and AdaBoost classifiers to enhance detection accuracy. Our methodology involves preprocessing the CIC-DDoS2019 dataset, extracting relevant features, and implementing both binary classification (benign vs. attack) and multiclass classification (attack type identification). The experimental results show that our ensemble model achieves an F1-score of 0.9967 for binary classification, with Gradient Boosting performing best among individual models. The multiclass classification reaches an accuracy of 0.8742 in distinguishing between different types of DDoS attacks. This research demonstrates that ensemble learning significantly improves the accuracy and reliability of DDoS attack detection compared to single-model approaches.

Keywords: Ensemble learning; DDoS attack; network security; CIC-DDoS2019; machine learning.

Analisis Kualitas Citra Steganografi Berbasis Spasial Pada Metode Least Significant Bits dan Pixel Value Differencing

Nizar Haris Masruri — Sat, 06 Sep 2025 00:00:00 +0000

Steganography is a method in network security systems that functions to hide confidential information by inserting it into other media without changing the authenticity of the information. One of the steganography media is images. Image steganography has two domains: spatial and frequency. The spatial domain works on the pixels of an image, and the frequency domain works on the image frequency. Spatial image steganography is said to be good if the level of stego-image damage is very small. There are several intermediate methods in spatial steganography, namely: LSB, PVD, GLM, EBE, RPE, and others. However, there are two most popular methods, namely LSB (Least Significant Bit) and PVD (Pixel Value Differencing). Therefore, this study aims to test the quality of stego-images produced by these two methods. Testing is carried out using the MSE and PSNR value parameters. A stego-image is considered good when the MSE value is close to 0 and the PSNR value is above 30 dB. From the test data used, this study produced an MSE value of 0.0006 for LSB stego-images and 0.007 for PVD stego-images. The PSNR value was 90 dB for LSB stego-images and 72 dB for PVD stego-images. This study concluded that LSB stego-images have less damage than PVD stego-images, so the researchers recommend using the LSB method in carrying out the steganography process.

Keywords: Spatial;MSE;PSNR;LSB;PVD

Perbandingan Kinerja Algoritma Naïve Bayes dan K-Nearest Neighbor dalam Menganalisis Sentimen Pengguna Game Free Fire

Nyoman Dinda Indira Sudiasta Putri, I Made Dendi Maysanjaya, I Made Gede Sunarya — Mon, 17 Nov 2025 00:00:00 +0000

Free Fire is one of the most popular online games in Indonesia, yet it continues to receive a wide range of user reviews regarding gameplay experiences. These reviews reflect diverse user perceptions, including both praise and criticism, making sentiment analysis essential to understanding user satisfaction. This study aims to classify user sentiments toward Free Fire using a combined dataset collected from the Google Play Store and App Store, and to compare the performance of two text classification algorithms: Naive Bayes and K-Nearest Neighbor (KNN). The data were collected using web scraping techniques and manually labeled by expert validators. Text preprocessing involved cleansing, tokenizing, stopword removal, and stemming, followed by term weighting using the Term Frequency-Inverse Document Frequency (TF-IDF) method. The experimental results show that the Naive Bayes algorithm achieved the highest accuracy of 72.78%, while the KNN algorithm recorded a maximum accuracy of 45.91%. Based on these findings, Naive Bayes is proven to be more effective in classifying user sentiments related to Free Fire. The results of this study are expected to provide constructive insights for developers to improve the quality and user experience of the game.

Penerapan Metode Dempster Shafer Pada Sistem Pakar Identifikasi Hama Tanaman Buah Naga Merah

Hanifah Nur Safitri, Desi Andreswari, Sempurna Br Ginting — Mon, 17 Nov 2025 00:00:00 +0000

Red dragon fruit is a type of fruit that has bioactive components such as flavonoids, phenolics, betacyanins and anthocyanins. Red dragon fruit has a distinctive combination of flavors, namely sweet, sour and refreshingly savory. Apart from that, red dragon fruit also has several health benefits so this fruit is liked by many people. However, a lack of knowledge in maintaining red dragon fruit from pest attacks can cause the growth of red dragon fruit to be less than optimal and result in crop failure. Therefore, an expert system for identifying red dragon fruit pests is needed to minimize the risk of crop failure. The use of an expert system in identifying red dragon fruit pests by applying the Dempster Shafer method can determine the appropriate treatment thereby reducing the percentage of crop failure. The Dempster Shafer method is used to determine the level of certainty of a symptom and provide an accurate level of confidence. The results of this expert system test show that the identification results are close to the truth from an expert using 9 types of pests and 24 symptoms. The resulting accuracy level has a percentage of 100%.

Keywords: Red Dragon Fruit, Expert System, Dempster Shafer Method

Model Penentuan Nilai Mahasiswa Pada Aspek Partisipasi Belajar Dengan Pendekatan Fuzzy Tsukamoto

Novriyanti Kesuma Kesuma, Armansyah Armansyah, Suhardi Suhardi — Mon, 17 Nov 2025 00:00:00 +0000

This study aims to determine the model of student learning participation value and also to determine the criteria for student participation that triggers the value of student learning participation. This study has 5 variables, namely attendance participation, questioning participation, giving ideas participation, helping friends and group discussion activity, each of which has a certain weight. The application of fuzzy logic in this assessment uses fuzzy tsukamoto, the fuzzy rules obtained are 46 rules. The defuzzification results get 87.47 with a very good category, so out of 31 participants who get a fairly good category are 21 participants, participants who get a good category are 9 participants and those who get a very good category are 1 participant.