MATRIK : Jurnal Manajemen, Teknik Informatika dan Rekayasa Komputer

Hostage Liberation Operations using Wheeled Robots Based on LIDAR (Light Detection and Ranging) Sensors

Sat, 13 Jan 2024 08:37:38 +0800

Hostage release operations require a high level of precision, alertness, and skill, which are carried out manually by soldiers of the Indonesian National Army. This medium presents a significant risk to soldiers. This research aims to improve the effectiveness of hostage release operations by integrating wheeled robot technology based on Light Detection and Ranging (LIDAR) sensors. The research method used is experiment-based in developing and testing a prototype of a mobile robot equipped with LIDAR technology and a web camera capable of mapping the location of hostages in three dimensions. The research showed that this robot has high accuracy, reaching 97.87%, and can create
three-dimensional route maps and display real-time video on a computer. The use of this technology has the potential to reduce risks to soldiers and improve the accuracy of mapping hostage locations, which can ultimately improve the safety and effectiveness of hostage release operations in the context of special operations tasks by soldiers of the Army.

Comparing Long Short-Term Memory and Random Forest Accuracy for Bitcoin Price Forecasting

Munirul Ula, Veri Ilhadi, Zailani Mohamed Sidek — Tue, 30 Jan 2024 09:11:30 +0800

Bitcoin’s daily value fluctuations are very dynamic. Understanding its rapid and intricate price movements demands advanced techniques for processing complex data. This research aims to compare the accuracy of two machine learning methods, Random Forest (RF) and Long Short-Term Memory (LSTM), in predicting Bitcoin price. This research employs RF and LSTM algorithms to forecast Bitcoin prices using a two-year Yahoo Finance dataset. The evaluation metrics used were accuracy based on Mean Absolute Percentage Error (MAPE) and computational power (CPU-Z). As a result of this research, the LSTM model demonstrates higher accuracy compared to the RF model. MAPE reveals LSTM’s precision of 99.8% and RF’s accuracy of 90.1%. Regarding computational time and resources, RF shows slightly better performance than LSTM. The visual comparison further emphasizes LSTM’s better performance in predicting Bitcoin prices, highlighting its potential for informed decision-making in cryptocurrency trading. This research contributes valuable insights into the effectiveness, strengths, and weaknesses of LSTM and RF models in predicting cryptocurrency trends.

Power Efficiency using Bank Capacitor Regulator on Field Service Shoes with Fast Charge Method

Fri, 09 Feb 2024 08:21:05 +0800

Power efficiency is a key factor in military equipment, including field service boots used by personnel in various field situations that often demand durability and reliable electricity availability. This research focused on improving the power efficiency of field service shoes by using capacitor bank regulators and fast charging methods. By designing and implementing this system, this research aims to optimize the use of power sources, extend battery life, and improve personnel comfort in the field. The method used in this research is the fast charge method. The fast charge method enables faster battery charging, which is important in field situations with limited time availability. The findings of this research show that the capacitor bank regulator can keep the DC output stable despite instability in the input. The total power usage in the circuit is 0.20 W, and the power efficiency is about 60.61%. The research shows the potential of this voltage conversion circuit for efficient applications. Although it has not achieved maximum efficiency, the capacitor bank regulator can maintain output stability even in input voltage instability. This circuit can effectively cope with voltage conversion in various applications with further optimization.

Regional Clustering Based on Types of Non-Communicable Diseases Using k-Means Algorithm

Tb Ai Munandar, Ajif Yunizar Yusuf Pratama — Fri, 09 Feb 2024 00:00:00 +0800

Noncommunicable diseases (NCDs) have become a global threat to public health, necessitating a comprehensive understanding of their geographic and epidemiological distribution in order to devise appropriate interventions. The objective of this study is to clustering areas of Banten Province based on NCDS profiles using the unsupervised learning technique. The method used in this study is the k-means algorithm for grouping types of non-communicable diseases based on region. The processing and normalisation of NCDS prevalence data from various health sources preceded cluster analysis using the k-means clustering algorithm. This research is categorised into two scenarios: the first involves the clustering of data obtained from outlier analysis, while the second scenario excludes any outliers. The objective is to observe disparities in regional clustering outcomes by categorising non-communicable diseases according to these two scenarios. The silhouette index is used to determine the validity of cluster results. These findings are analysed in depth to determine the geographic and socioeconomic patterns associated with each cluster's NCDS profile. Based on the mean silhouette index value of 0.812, the results indicate that the sum of k = 2 in the k-means algorithm is the optimal cluster result in this case. Five non-communicable diseases, namely diabetes, hypertension, obesity, stroke, and cataracts, necessitate significant focus in the first cluster (C1), where 202 regions were grouped. Six regions belong to the second cluster (C2), which includes areas that are not only susceptible to the five non-communicable diseases in cluster C1 but also to breast cancer, cervical cancer, heart disease, chronic obstructive pulmonary disease (COPD), and congenital deafness.

Learning Accuracy with Particle Swarm Optimization for Music Genre Classification Using Recurrent Neural Networks

Muhammad Rizki, Arief Hermawan, Donny Avianto — Tue, 20 Feb 2024 15:19:52 +0800

Deep learning has revolutionized many fields, but its success often depends on optimal selection hyperparameters, this research aims to compare two sets of learning rates, namely the learning set rates from previous research and rates optimized for Particle Swarm Optimization. Particle Swarm Optimization is learned by mimicking the collective foraging behavior of a swarm of particles, and repeatedly adjusting to improve performance. The results show that the level of Particle Swarm Optimization is better previous level, achieving the highest accuracy of 0.955 compared to the previous best accuracy level of 0.933. In particular, specific levels generated by Particle Swarm Optimization, for example, 0.00163064, achieving competitive accuracy of 0.942-0.945 with shorter computing time compared to the previous rate. These findings underscore the importance of choosing the right learning rate for optimizing the accuracy of Recurrent Neural Networks and demonstrating the potential of Particle Swarm Optimization to exceed existing research benchmarks. Future work will explore comparative analysis different optimization algorithms to obtain the learning rate and assess their computational efficiency. These further investigations promise to improve the performance optimization of Recurrent Neural Networks goes beyond the limitations of previous research.

Modeling the Farmer Exchange Rate in Indonesia Using the Vector Error Correction Model Method

Yuniar Farida, Afanin Hamidah, Silvia Kartika Sari, Lutfi Hakim — Wed, 21 Feb 2024 14:17:08 +0800

The agricultural sector plays a crucial role in the Indonesian economy. However, the farm sector still has serious problems, including agricultural product prices, which often fall when the harvest supply is abundant. So often, the income obtained is not proportional to the price spent by farmers, which has an impact on decreasing the welfare of farmers. An indicator to observe changes in the interest of Indonesian farmers is the Farmer Exchange Rate Index (NTP). This study aims to form a model and project the welfare level of farmers in Indonesia, focusing on NTP indicators, which are caused by the influence of variables such as inflation, Gross Domestic Product (GDP), interest rates, and the rupiah exchange rate. The method used is the Vector Error Correction Model (VECM), used when there are indications that the research variables do not show stability at the initial level and there is a cointegration relationship. The results of this study show that in the long run, significant factors affecting NTP are inflation, interest rates, and the rupiah exchange rate. Meanwhile, in the short term, the variables that have an impact are GDP and the rupiah exchange rate. The resulting VECM model shows a MAPE error rate of 1.79%, indicating excellent performance, as the MAPE error rate is below 10%. The implication of this research is provides information related to NTP projection that can be used to formulate strategies to strengthen Indonesia's agricultural sector.

Optimizing Treatment of Herbal Plant Using SOPHERBAL Android Application Fordward Chaining Method

Muhamad Azwar, Eka Nurul Qomaliyah, Nurul Indriani — Tue, 27 Feb 2024 16:36:33 +0800

The utilization of traditional herbal medicine among the inhabitants of Lombok is notably prevalent yet frequently hindered by a lack of comprehension regarding the efficacy of herbal remedies for specific ailments. Addressing this challenge, this study proposes the development of an Android application called ”sopherbal,” aimed at delivering personalized herbal plant recommendations via easily accessible mobile devices. Employing forward chaining methodology, the application identifies optimal herbal remedies based on ailment type, processing techniques, usage instructions, and recommended dosage and treatment duration. Notably, while effective in this context, the forward chaining approach entails certain trade-offs and hurdles. Previous research indicates that forward chaining facilitates
accurate recommendation generation, and it may be constrained by its reliance on predefined rules and limited adaptability to complex, evolving scenarios. Despite these challenges, the ”sopherbal” application, featuring 50 Sasak medicinal plants curated for 15 common ailments, achieved an 86% validation rate, affirming its efficacy in bridging the gap between traditional herbal knowledge and modern healthcare needs.

Comparison of DenseNet-121 and MobileNet for Coral Reef Classification

Heru Pramono Hadi, Eko Hari Rachmawanto, Rabei Raad Ali — Fri, 08 Mar 2024 09:21:49 +0800

Coral reefs are a type of marine organism that has beauty and benefits for other sea creatures’ ecosystems. However, despite its beauty and usefulness, coral reefs are vulnerable to damage such as coral bleaching, which can impact other coral reef ecosystems. This research aims to classify digital images of healthy, bleached, and dead coral reefs. This research method is DenseNet-121 and MobileNet is based on Convolutional Neural Networks. This research uses a dataset from 1582 coral reef image data with three main classes: 720 were bleached, 150 were dead, and 712 were healthy. The testing process is carried out using several forms of split datasets, namely 60:10:30, 50:10:40, and 70:10:20. The test results obtained with a data sharing percentage of 60:10:30 show that MobileNet architecture achieved 88.00% accuracy, and DenseNet-121 achieved 91.57% accuracy. Using a data split percentage of 50:10:40, MobileNet achieved 84.51% accuracy, and DenseNet- 121 achieved 90.52% accuracy. Meanwhile, with a data separation percentage of 70:10:20, MobileNet achieved 85.48% accuracy, and DenseNet-121 achieved 92.74% accuracy.

Improving Performance Convolutional Neural Networks Using Modified Pooling Function

Achmad Lukman, Wahju Tjahjo Saputro, Erni Seniwati — Fri, 08 Mar 2024 09:43:26 +0800

The Visual Geometry Group-16 (VGG16) network architecture, as part of the development of convolutional neural networks, has been popular among researchers in solving classification tasks, so in this paper, we investigated the number of layers to find better performance. In addition, we also proposed two pooling function techniques inspired by existing research on mixed pooling functions, namely Qmax and Qavg. The purpose of the research was to see the advantages of our method; we conducted several test scenarios, including comparing several modified network configurations based on VGG16 as a baseline and involving our pooling technique and existing pooling functions. Then, the results of the first scenario, we selected a network that can adapt well to our pooling technique, which
was then carried out several tests involving the Cifar10, Cifar100, TinyImageNet, and Street View House Numbers (SVHN) datasets as benchmarks. In addition, we were also involved in several existing methods. The experiment results showed that Net-E has the highest performance, with 93.90% accuracy for Cifar10, 71.17% for Cifar100, and 52.84% for TinyImageNet. Still, the accuracy was low when the SVHN dataset was used. In addition, in comparison tests with several optimization algorithms using the Qavg pooling function, it can be seen that the best accuracy results lie in the SGD optimization algorithm, with 89.76% for Cifar10 and 89.06% for Cifar100.

Enhancing Predictive Models: An In-depth Analysis of Feature Selection Techniques Coupled with Boosting Algorithms

Neny Sulistianingsih, Galih Hendro Martono — Fri, 08 Mar 2024 10:59:05 +0800

This research addresses the critical need to enhance predictive models for fetal health classification using Cardiotocography (CTG) data. The literature review underscores challenges in imbalanced labels, feature selection, and efficient data handling. This paper aims to enhance predictive models for fetal health classification using Cardiotocography (CTG) data by addressing challenges related to imbalanced labels, feature selection, and efficient data handling. The study uses Recursive Feature Elimination (RFE) and boosting algorithms (XGBoost, AdaBoost, LightGBM, CATBoost, and Histogram-Based Boosting) to refine model performance. The results reveal notable variations in precision, Recall, F1-Score, accuracy, and AUC across different algorithms and RFE applications. Notably, Random Forest with XGBoost exhibits superior performance in precision (0.940), Recall (0.890), F1-Score (0.920), accuracy (0.950), and AUC (0.960). Conversely, Logistic Regression with AdaBoost demonstrates lower performance. The absence of RFE also impacts model effectiveness. In conclusion, the study successfully employs RFE and boosting algorithms to enhance fetal health classification models, contributing valuable insights for improved prenatal diagnosis.

Detecting Hidden Illegal Online Gambling on .go.id Domains Using Web Scraping Algorithms

Muchlis Nurseno, Umar Aditiawarman, Haris Al Qodri Maarif, Teddy Mantoro — Fri, 08 Mar 2024 00:00:00 +0800

The profitable gambling business has encouraged operators to promote online gambling using black hat SEO by targeting official sites such as government sites. Operators have used various techniques to prevent search engines from distinguishing between genuine and illegal content. This research aims to determine whether websites with the go.id domain have been compromised with hidden URLs affiliated with online gambling sites. The method used in this research is an experiment using a FOFA.info dataset containing a complete list of 450,000 .go.id domains. A web scraping algorithm developed in Python was used to identify potentially compromised websites from the targeted list
by analyzing gambling-related keywords in local languages, such as ’slot,’ ’judi,’ ’gacor,’ and ’togel'. The results showed that 958 of the 1,482 suspected.go.id sites had been compromised with an accuracy rate of 99.1%. This implies that security gaps have been exploited by illegal online gambling sites, posing a reputational risk to the government. Lastly, the scrapping algorithm tool developed in this research can detect illegal online gambling hidden in domains such as .ac.id, .or.id, .sch.id, and help authorities take necessary action.

Educational Data Mining: Multiple Choice Question Classification in Vocational School

Sucipto Sucipto, Didik Dwi Prasetya, Triyanna Widiyaningtyas — Sat, 16 Mar 2024 00:00:00 +0800

Data mining on student learning outcomes in the education sector can overcome this problem. This research aimed to provide a solution for selecting quality multiple choice questions (MCQ) using the results of students’ mid-semester exams in vocational high schools using a Data Mining approach. The research method used was the Cross-Industry Standard Process for Machine Learning (CRISP-ML) model. Steps to assess the accuracy of analyzing the difficulty level of questions based on student profile data and midterm test results. The data used in this research were the findings of basic computer tests on mid-term exams in mathematics disciplines at vocational high schools. This research used several classification algorithms, including SVM, Naive Bayes, Random Forest, Decision Three, Linear Regression, and KNN. The results of evaluating the classification

Unsafe Conditions Identification Using Social Networks in Power Plant Safety Reports

Annisa’ul Mubarokah, Rita Ambarwati, Dedy Dedy, Mashhura Toirхonovna Alimova — Sat, 16 Mar 2024 00:00:00 +0800

Power plants in Indonesia grapple with significant challenges in managing occupational health and safety. Power generation companies urgently need to reduce workplace accidents every year and need an application for reporting every potential workplace hazard. The huge reporting data in applications such as IZAT requires thorough analysis to find out the pattern and distribution. This research aims to facilitate the company in hazard mitigation by identifying reported unsafe conditions and building a semantic association network to understand the nature of unsafe conditions between Paiton and Indramayu generating units. The research method uses social network analysis, which is carried out by preprocessing the data using programming to remove noise and then converting the data into a readable format. Then, semantic relationships between words were analyzed, and the data was visualized using the ForceAtlas2 algorithm. The findings revealed a different focus between the two units, where 6.597 reports from the Paiton generating unit mainly highlighted team response and accident-prone workplace conditions, while 5.840 reports from the Indramayu unit emphasized specific conditions, locations, and equipment that pose accident risks

Enhancing Accuracy in Stock Price Prediction: The Power of Optimization Algorithms

Vivi Aida Fitria, Lilis Widayanti — Mon, 25 Mar 2024 00:00:00 +0800

The purpose of this research was to improve the accuracy of stock price prediction by implementing optimization algorithms on forecasting methods, in this case, the exponential smoothing method. This research implemented the Particle Swarm Optimization (PSO) and Bat Algorithm metaheuristic optimization algorithms to determine the single-exponential smoothing method’s smoothing parameters. Before implementing the optimization algorithm, the way to determine the smoothing parameters was by trial-and-error method, which is considered less effective. Therefore, the novelty of this research is tuning the parameters of the exponential smoothing method using a comparison of two metaheuristic algorithms, namely the particle swarm optimization algorithm compared to the bat algorithm. The Single Exponential Smoothing method with PSO and Bat algorithms was proven to improve accuracy. The alpha parameter found by the PSO algorithm is 0.9346, and the bat algorithm is 0.936465. With a MAPE of 1.0311%, it was better than the MAPE generated in the Single Exponential smoothing method by trial and error of 1.0316%. This research contributes to providing insight that in a highly sensitive stock prediction situation, metaheuristic algorithms can be used to create more accurate and efficient prediction results.

Optimization of SVM and Gradient Boosting Models Using GridSearchCV in Detecting Fake Job Postings

Tue, 26 Mar 2024 00:00:00 +0800

Online job searching is one of the most efficient ways to do this, and it is widely used by people worldwide because of the automated process of transferring job recruitment information. The easy and fast process of transferring information in job recruitment has led to the rise of fake job vacancy fraud. Several studies have been conducted to predict fake job vacancies, focusing on improving accuracy. However, the main problem in prediction is choosing the wrong parameters so that the classification algorithm does not work optimally. This research aimed to increase the accuracy of fake job vacancy predictions by tuning parameters using GridSearchCV. The research method used was SVM and Gradient Boosting with parameter adjustments to improve the parameter combination and align it with the predicted model characteristics. The research process was divided into preprocessing, feature extraction, data separation, and modeling stages. The model was tested using the EMSCAD dataset. This research showed that the SVM algorithm can achieve the highest accuracy of 98.88%, while gradient enhancement produces an accuracy of 98.08%. This research showed that optimizing the SVM model with GridSearchCV can increase accuracy in predicting fake job recruitment.

Forecasting the Poverty Rates using Holt’s Exponential Smoothing

Riza Prapascatama Agusdin, Sylvert Prian Tahalea, Vynska Amalia Permadi — Tue, 26 Mar 2024 00:00:00 +0800

As a developing country with many provinces, Indonesia has a poverty problem that needs to be overcome. This research aimed to predict the poverty level in the Special Region of Yogyakarta using poverty data provided by the Central Statistics Agency for the Special Region of Yogyakarta. The method used in this research was Holt exponential smoothing to predict poverty levels in Yogyakarta City and four districts (Sleman, Bantul, Kulon Progo, and Gunungkidul) in this province. Three performances were measured to evaluate forecast results: sum squared error, mean squared error, and root mean squared error. The research results showed that the best configuration for the cities of Yogyakarta and Bantul is , = 0.9, 0.4; Kulon Progo and Gunungkidul are , = 0.9, 0.9; and Sleman are , = 0.9, 0.6. The forecasting results for 2022 to 2024, using a 95% confidence interval, showed that the poverty rate will increase in every city and district in the Special Region of Yogyakarta.

Sentiment Analysis of e-Government Service Using the Naive Bayes Algorithm

Tue, 26 Mar 2024 00:00:00 +0800

A digital platform called the Alpukat Population Application is used to handle statistics and information regarding DKI Jakarta's population. Using the Naive Bayes Classifier (NBC) approach, sentiment analysis for applications using satellite placement. The Nave Bayes Classifier technique is utilized for sentiment analysis because of its benefits in modeling and categorizing complicated data. The user reviews and comments gathered from the Google Play Store were the source of the data utilized in this research. Feature extraction using methods like TF-IDF, sentiment labeling on data, and the development of Nave Bayes Classifier models for sentiment classification were all part of the research project. It is anticipated that the study's findings would help us better understand how users interact with the Alpukat population app. This sentiment analysis may assist app administrators and developers in identifying the positives and negatives of applications and planning updates and advancements based on user feedback. It is anticipated that the sentiment classification model created using the Naive Bayes Classifier approach would be able to classify user evaluations into positive, negative, or neutral sentiment categories with a high degree of accuracy. The creation of improved alpukat positioning apps and decision-making may both benefit from this emotive analysis.

DenseNet Architecture for Efficient and Accurate Recognition of Javanese Script Hanacaraka Character

Egi Dio Bagus Sudewo, Muhammad Kunta Biddinika, Abdul Fadlil — Thu, 28 Mar 2024 00:00:00 +0800

This study introduced a specifically optimized DenseNet architecture for recognizing Javanese Hanacaraka characters, focusing on enhancing efficiency and accuracy. The research aimed to preserve and celebrate Java’s rich cultural heritage and historical significance through the development of precise character recognition technology. The method used advanced techniques within convolutional neural networks (CNN) to integrate feature extraction across densely connected layers efficiently. The result of this study was that the developed model achieved a training accuracy of 100% and a validation accuracy of approximately 99.50% after 30 training epochs. Furthermore, when tested on previously unseen datasets, the model exhibited exceptional accuracy, precision, recall, and F1-score, reaching 100%. These findings underscored the remarkable capability of DenseNet architecture in character recognition, even across novel datasets, suggesting significant potential for automating Javanese Hanacaraka text processing across various applications, ranging from text recognition to digital archiving. The conclusion drawn from this study suggests that optimizing DenseNet architecture can be a significant step in preserving and developing character recognition technology for Javanese

Implementation of Neural Machine Translation in Translating from Indonesian to Sasak Language

Helna Wardhana, I Made Yadi Dharma, Khairan Marzuki, Ibjan Syarif Hidayatullah — Thu, 28 Mar 2024 00:00:00 +0800

Language translation is part of Natural Language Processing, also known as Machine Translation, which helps the process of learning foreign and regional languages using translation technology in sentence form. In Lombok, there are still people who are not very fluent in Indonesian because Indonesian is generally only used at formal events. This research aimed to develop a translation model from Indonesian to Sasak. The method used was the Neural Machine Translation with the Recurrent Neural Network - Long Short Term Memory architecture and the Word2Vec Embedding with a sentence translation system. The dataset used was a parallel corpus from the Tatoeba Project and other open sources, divided into 80% training and 20% validation data. The result of this research was the application of Neural Machine Translation with the Recurrent Neural Network - Long Short Term Memory algorithm, which could produce a model with an accuracy of 99.6% in training data and 71.9% in test data. The highest ROUGE evaluation metric result obtained on the model was 88%. This research contributed to providing a translation model from Indonesian to Sasak for the local community to facilitate communication and preserve regional language culture.

Stroke Prediction with Enhanced Gradient Boosting Classifier and Strategic Hyperparameter

Sat, 30 Mar 2024 00:00:00 +0800

A stroke is a medical condition that occurs when the blood supply to the brain is interrupted. Stroke can cause damage to the brain that can potentially affect a person's function and ability to move, speak, think, and feel normally. The effect of stroke on health emphasizes the importance of stroke detection, so an effective model is needed in predicting stroke. This research aimed to find a new approach that can improve the performance of stroke prediction by comparing four derivative algorithms from Gradient Boosting by adding hyperparameters tuning. The addition of hyperparameters was used to find the best combination of parameter values that can improve the model accuracy. The methods used in this research were Categorical Boosting, Histogram Gradient Boosting, Light Gradient Boosting, and Extreme Gradient Boosting. The research involved retrieving, cleaning, and analyzing data and then the model performance was evaluated with a confusion matrix and execution time. The results obtained were Light Gradient Boosting with Hyperparameter RandomSearchCV achieved the highest accuracy at 95% among the algorithms tested, while also being the fastest in execution. The contribution of this research to the medical field can help doctors and patients predict the occurrence of stroke early and reduce serious consequences.