Article Per Year (5 Year)
p-Index From 2020 - 2025
0.408
P-Index
This Author published in this journals
All Journal CogITo Smart Journal Journal of Informatics, Information System, Software Engineering and Applications (INISTA)
Kusuma, Adi Wahyu Candra
Unknown Affiliation
Computer Science & IT Decision Sciences, Operations Research & Management Education Electrical & Electronics Engineering

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 2 Documents
Search

 1 
Analysis Comparison of K-Nearest Neighbor, Multi-Layer Perceptron, and Decision Tree Algorithms in Diamond Price Prediction Kamila, Ahya Radiatul; Andry, Johanes Fernandes; Kusuma, Adi Wahyu Candra; Prasetyo, Eko Wahyu; Derhass, Gerry Hudera
CogITo Smart Journal Vol. 10 No. 2 (2024): Cogito Smart Journal
Publisher : Fakultas Ilmu Komputer, Universitas Klabat

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31154/cogito.v10i2.532.298-311

Abstract

Diamond price predictions are essential due to the high demand for these gemstones, valued as investments and jewelry. Diamonds are expensive due to their rarity and extraction process. Their prices vary depending on key factors like the diamond's inherent value and secondary factors such as marketing costs, brand names, and market trends. These variations often confuse customers, potentially leading to investment losses. This research aims to help investors determine the true price of diamonds based solely on their intrinsic value, excluding secondary factors. A machine learning approach was utilized to predict diamond prices, focusing on primary determinants. Three models such as Multi-Layer Perceptron (MLP), Decision Tree, and K-Nearest Neighbor (KNN) were compared with manual hyperparameter tuning to identify the best performing algorithm. Model performance was evaluated using Mean Absolute Percentage Error (MAPE), Mean Absolute Error (MAE), and Mean Squared Error (MSE). Among the models, KNN demonstrated the best results, achieving MAPE, MAE, and MSE values of 1.1%, 0.00038, and 〖2.687 x 10〗^(-6) respectively. This study offers valuable insights for investors by accurately predicting diamond prices based on fundamental attributes, minimizing the impact of secondary factors.
Implementation of Random Forest Classification and Support Vector Machine Algorithms for Phishing Link Detection Tampinongkol, Felliks Feiters; Kamila, Ahya Radiatul; Wardhana, Ariq Cahya; Kusuma, Adi Wahyu Candra; Revaldo, Danny
Journal of INISTA Vol 7 No 1 (2024): November 2024
Publisher : LPPM Institut Teknologi Telkom Purwokerto

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.20895/inista.v7i1.1588

Abstract

This research compares two machine learning methods, Support Vector Machine (SVM) and Random Forest Classification (RFC), in detecting phishing links. Phishing is an attempt to obtain sensitive information by masquerading as a trustworthy entity in electronic communications. Detecting phishing links is crucial in protecting users from this cyber threat. In this study, we used a dataset consisting of features extracted from URLs, such as URL length, the use of special characters, and domain information. The dataset was then split into training and testing data with an 80:20 ratio. We trained the SVM and RFC models using the training data and evaluated their performance based on the testing data. The results show that both methods have their respective advantages. SVM, known for handling high-dimensional data well and providing optimal solutions for classification problems, demonstrated a high accuracy rate in detecting phishing links. However, SVM requires a longer training time compared to RFC. On the other hand, RFC, an ensemble method known for its resilience to overfitting, showed performance nearly comparable to SVM in terms of accuracy but with faster training time and better interpretability. This comparison indicates that RFC is more suitable for scenarios requiring quick results and easy interpretation, while SVM is more appropriate for situations where accuracy is critical, and computational resources are sufficient. In conclusion, the choice of phishing link detection method should be tailored to specific needs and available resource constraints. This research provides valuable insights for developing more effective, efficient, and relevant phishing detection systems.
Co-Authors Ariq Cahya Wardhana Derhass, Gerry Hudera EKO WAHYU PRASETYO Johanes Fernandes Andry Kamila, Ahya Radiatul Revaldo, Danny Tampinongkol, Felliks Feiters