Jurnal Online Informatika

Regression Analysis for Crop Production Using CLARANS Algorithm

Arie Vatresia, Ruvita Faurina, Yanti Simanjuntak — Wed, 28 Jun 2023 00:00:00 +0000

Crop production rate relies on rainfall over Rejang Lebong district. Data showed a discrepancy between increased crop production and rainfall in Rejang Lebong District. However, the

spatiotemporal distribution of the crop variable's dependencies remains unclear. This study analyses the relationship between rainfall and crop production rate in the Rejang Lebong district based on the performance of the machine learning method. In addition, this research also performed regression analysis to carry out rainfall clusters and crop production. This order provides information in the form of cluster results to determine how much the rainfall variable influences the crop production rate in each cluster. Harnessing the Elbow, CLARANS, Simple Linear Regression, and Silhouette Coefficient methods, this study used 231 rainfall data sourced from the Bengkulu BMKG and 110 data for plant production obtained from BPS Bengkulu Province from 2000-2022. This research found that the optimal clusters were 3 clusters. C₁ contains 106 data with the largest regression value for chili = 0.127, C₂ contains 15 data with the largest regression value for mustard greens = 0.135, and C₃ contains 110 data with the largest regression value for cabbage = 0.408, eggplant = 0.197, and carrots = 0.201. Furthermore, this research also found that the biggest correlation of crops with highly significant improvement would be cabbage commodity (Y=0.4114X+0.2013) and chili plantation with high RSME (0.9897).

Scalability Testing of Land Forest Fire Patrol Information Systems

Ahmad Khusaeri, Imas Sukaesih Sitanggang, Hendra Rahmawan — Wed, 28 Jun 2023 00:00:00 +0000

The Patrol Information System for the Prevention of Forest Land Fires (SIPP Karhutla) in Indonesia is a tool for assisting patrol activities for controlling forest and land fires in Indonesia. The addition of Karhutla SIPP users causes the need for system scalability testing. This study aims to perform non-functional testing that focuses on scalability testing. The steps in scalability testing include creating schemas, conducting tests, and analyzing results. There are five schemes with a total sample of 700 samples. Testing was carried out using the JMeter automation testing tool assisted by Blazemeter in creating scripts. The scalability test parameter has three parameters: average CPU usage, memory usage, and network usage. The test results show that the CPU capacity used can handle up to 700 users, while with a memory capacity of 8GB it can handle up to 420 users. All users is the user menu that has the highest value for each test parameter The average value of CPU usage is 44.8%, the average memory usage is 69.48% and the average network usage is 2.8 Mb/s. In minimizing server performance, the tile cache map method can be applied to the system and can increase the memory capacity used.

Social Network Analysis: Identification of Communication and Information Dissemination (Case Study of Holywings)

Umar Aditiawarman, Mega Lumbia, Teddy Mantoro, Adamu Abubakar Ibrahim — Wed, 28 Jun 2023 00:00:00 +0000

Social media especially Twitter has been used by corporation or organization as an effective tool to interact and communicate with the consumers. Holywings is one of the popular restaurants in Indonesia that use social media as a tool to promote and disseminate information regarding their products and services. However, one of their promotional items has gone viral and invited public protests which turned into a trending topic on Twitter for a couple of weeks. Holywings allegedly improperly promoted their products by using the most honorable names, “Muhammad” and “Maria”. Social network analysis of Twitter data is conducted to identify and examine information circulating among the users, which leads to wider public attention and law enforcement. In this study, we focused on the conversation about Holywings on Twitter from 24 June to 31 July 2022. The analysis was carried out using Python to retrieve data and Gephi software to visualize the interactions and the intensity of the network group in viewing the spread of information. The findings reveal the centrality account that caused the news to go viral are the CNN Indonesia (@CNNIndonesia) news media account and Haris Pertama (@knpiharis), with a centrality of 0.161 and 0.282, respectively. There are also 121 groups involved in the conversation with modularity of 0.821.

Run Length Encoding Compresion on Virtual Tour Campus to Enhance Load Access Performance

Ade Bastian, Ardi Mardiana, Mega Berliani, Mochammad Bagasnanda Firmansyah — Wed, 28 Jun 2023 00:00:00 +0000

Virtual tour is one of the rapidly growing applications of multimedia technology which is used for various purposes, including the dissemination of information in an interesting way. The education sector is also not spared from using virtual tour media for promotional purposes, and campuses are no exception to this rule. Large virtual tour content causes high access speed, ultimately reducing the level of comfort experienced by users. This study aims to compress panoramic images displayed on a campus virtual tour using a lossless compression method and the Run Length Encoding (RLE) algorithm. First, panoramic images are combined into one, then individual images are compressed. When recreating a virtual campus tour, compressed images are used so that the amount of data transferred is smaller. The load access speed index increases from 7,233 seconds to 3,789 seconds when images are compressed from 64 bits to 8 bits, with a compression percentage of 27%. The findings from this research are that the RLE algorithm has not been able to compress large files effectively even though it is quite successful in increasing the load access of the virtual tour website.

The Implementation of Restricted Boltzmann Machine in Choosing a Specialization for Informatics Students

Vinna Rahmayanti Setyaning Nastiti, Zamah Sari, Bella Chintia Eka Merita — Wed, 28 Jun 2023 00:00:00 +0000

Choosing a specialization was not an easy task for some students, especially for those who lacked confidence in their skill and ability. Specialization in tertiary education became the benchmark and key to success for students’ future careers. This study was conducted to provide the learning outcomes record, which showed the specialization classification for the Informatics students by using the data from the students of 2013-2015 who had graduated. The total data was 319 students. The classification method used for this study was the Restricted Boltzmann Machine (RBM). However, the data showed imbalanced class distribution because the number of each field differed greatly. Therefore, SMOTE was added to classify the imbalanced class. The accuracy obtained from the combination of RBM and SMOTE was 70% with a 0.4 mean squared error.

Implementation of Generative Adversarial Network to Generate Fake Face Image

Jasman Pardede, Anisa Putri Setyaningrum — Wed, 28 Jun 2023 00:00:00 +0000

In recent years, many crimes use technology to generate someone's face which has a bad effect on that person. Generative adversarial network is a method to generate fake images using discriminators and generators. Conventional GAN involved binary cross entropy loss for discriminator training to classify original image from dataset and fake image that generated from generator. However, use of binary cross entropy loss cannot provided gradient information to generator in creating a good fake image. When generator creates a fake image, discriminator only gives a little feedback (gradient information) to generator update its model. It causes generator take a long time to update the model. To solve this problem, there is an LSGAN that used a loss function (least squared loss). Discriminator can provide a
strong gradient signal to generator update the model even though image was far from decision boundary. In making fake images, researchers used Least Squares GAN (LSGAN) with discriminator-1 loss value is 0.0061, discriminator-2 loss value is 0.0036, and generator loss value is 0.575. With the small loss value of the three important components, discriminator accuracy value in terms of classification reaches 95% for original image and 99% for fake image. In classified original image and fake image in this study
using a supervised contrastive loss classification model with an accuracy value of 99.93%.

Catbreedsnet: An Android Application for Cat Breed Classification Using Convolutional Neural Networks

Anugrah Tri Ramadhan, Abas Setiawan — Wed, 28 Jun 2023 00:00:00 +0000

There are so many cat races in the world. Ignorance in recognizing cat breeds will be dangerous if the cat being kept is affected by a disease, which allows mishandling of the cat being kept. In addition, many cat breeds have different foods from one race to another. The problem is that a cat caretaker cannot easily recognize the cat breed. Therefore, technology needs to help a cat caretaker to treat cats appropriately. In this study, we proposed a Machine Learning approach to recognize cat breeds. This study aims to identify the cat breed from the cat images then deployed on an Android smartphone. It was tested with data from cat images of 13 races. The classification method applied in this study uses the Convolutional Neural Network (CNN) algorithm using transfer learning. The base models tested are MobilenetV2, VGG16, and InceptionV3. The results tested using several models and through several experimental scenarios produced the best classification model with an accuracy of 82% with MobilenetV2. The model with the best accuracy is then embedded in an application with the Android operating system. Then the application is named Catbreednet.

Malware Image Classification Using Deep Learning InceptionResNet-V2 and VGG-16 Method

Didih Rizki Chandranegara, Jafar Shodiq Djawas, Faiq Azmi Nurfaizi, Zamah Sari — Wed, 28 Jun 2023 00:00:00 +0000

Malware is intentionally designed to damage computers, servers, clients or computer networks. Malware is a general term used to describe any program designed to harm a computer or server. The goal is to commit a crime, such as gaining unauthorized access to a particular system, so as to compromise user security. Most malware still uses the same code to produce another different form of malware variants. Therefore, the ability to classify similar malware variant characteristics into malware families is a good strategy to stop malware. The research is useful for classifying malware on malware samples presented as bytemap grayscale images. The malware classification research focused on 25 malware classes with a total of 9,029 images from the Malimg dataset. This research implements the VGG-16 and InceptionResNet-V2 architectures by running 2 different scenarios, scenario 1 uses the original dataset and the other scenario uses the undersampled dataset. After building the model, each scenario will get an evaluation form such as accuracy, precision, recall, and f1-score. The highest score was obtained in scenario 2 on the VGG-16 method with a score of 94.8% and the lowest in scenario 2 on the InceptionResNet-V2 method with a score of 85.1%.

Comparative Analysis of Machine Learning-based Forest Fire Characteristics in Sumatra and Borneo

Ayu Shabrina, Intan Nuni Wahyuni, Arnida L Latifah — Wed, 28 Jun 2023 00:00:00 +0000

Sumatra and Borneo are areas consisting of rainforests with a high vulnerability to fire. Both areas are in the tropics which experience rainy and dry seasons annually. The long dry season such as in 2019 triggered forest and land fires in Borneo and Sumatra, causing haze disasters in the exposed areas. This indicates that climate variables play a role in burning forests and land in Borneo and Sumatra, but how climate affects the fires in both areas is still questionable. This study investigates the climate variables: temperature, humidity, precipitation, and wind speed in relation to the fire’s characteristics in Borneo and Sumatra. We use the Random Forest model to determine the characteristics of forest fires in Sumatra and Borneo based on the climate variables and carbon emission levels. According to the model, the fire event in Sumatra is slightly better predicted than in Borneo, indicating a climate-fire dependence is more prominent in Sumatra. Nevertheless, a maximum temperature variable is seemingly an important indicator for forest and land fire in both domains as it gives the largest contribution to the carbon emission.

Implementation of Ant Colony Optimization – Artificial Neural Network in Predicting the Activity of Indenopyrazole Derivative as Anti-Cancer Agent

Isman Kurniawan, Nabilla Kamil, Annisa Aditsania, Erwin Budi Setiawan — Wed, 28 Jun 2023 00:00:00 +0000

Cancer is a disease induced by the abnormal growth of cells in body tissues. This disease is commonly treated by chemotherapy. However, at first, cancer cells can respond to the activity of chemotherapy over time, but over time, resistance to cancer cells appears. Therefore, it is required to develop new anti-cancer drugs. Indenopyrazole and its derivative have been investigated to be a potential drug to treat cancer. This study aims to predict indenopyrazole derivative compounds as anti-cancer drugs by using Ant Colony Optimization (ACO) and Artificial Neural Network (ANN) methods. We used 93 compounds of indenopyrazole derivative with a total of 1876 descriptors. Then, the descriptors were reduced by using the Pearson Correlation Coefficient (PCC) and followed by the ACO algorithm to get the most relevant features. We found that the best number of descriptors obtained from ACO is ten descriptors. The ANN prediction model was developed with three architectures, which are different in hidden layer number, i.e., 1, 2, and 3 hidden layers. Based on the results, we found that the model with three hidden layers gives the best performance, with the value of the R2 test, R2 train, and Q2 train being 0.8822, 0.8495, and 0.8472, respectively.

Data Mining for Heart Disease Prediction Based on Echocardiogram and Electrocardiogram Data

Tb Ai Munandar — Wed, 28 Jun 2023 00:00:00 +0000

Traditional methods of detecting cardiac illness are often problematic in the medical field. The doctor must next study and interpret the findings of the patient's medical record received from the electrocardiogram and echocardiogram. These tasks often take a long time and require patience. The use of computational technology in medicine, especially the study of cardiac disease, is not new. Scientists are continuously striving for the most reliable method of diagnosing a patient's cardiac illness, particularly when an integrated system is constructed. The study attempted to propose an alternative for identifying cardiac illness using a supervised learning technique, namely the multi-layer perceptron (MLP). The study started with the collection of patient medical record data, which yielded up to 534 data points, followed by pre-processing and transformation to provide up to 324 data points suitable to be employed by learning algorithms. The last step is to create a heart disease classification model with distinct activation functions using MLP. The degree of classification accuracy, k-fold cross-validation, and bootstrap are all used to test the model. According to the findings of the study, MLP with the Tanh activation function is a more accurate prediction model than logistics and Relu. The classification accuracy level (CA) for MLP with Tanh and k-fold cross-validation is 0.788 in a data-sharing situation, while it is 0.672 with Bootstrap. MLP using the Tanh activation function is the best model based on the CA level and the AUC value, with values of 0.832 (k-fold cross-validation) and 0.857 (bootstrap).

Classification of Stunting in Children Using the C4.5 Algorithm

Muhajir Yunus, Muhammad Kunta Biddinika, Abdul Fadlil — Wed, 28 Jun 2023 00:00:00 +0000

Stunting is a disease caused by malnutrition in children, which results in slow growth. Generally, stunting is characterized by a lack of weight and height in young children. This study aims to classify stunting in children aged 0-60 months using the Decision Tree C4.5 method based on z-score calculations with a sample size of 224 records, consisting of 4 attributes and 1 label, namely Gender, Age, Weight, Height, and Nutritional Status. The results of the study obtained a C4.5 decision tree where the Age variable influenced the classification of stunting with the highest Gain Ratio of 0.185016337. Meanwhile, the evaluation of the model using the Confusion matrix resulted in the highest accuracy of 61.82% and AUC of 0.584.

Multi-Step Vector Output Prediction of Time Series Using EMA LSTM

Mohammad Diqi; Ahmad Sahal, Farida Nur Aini — Wed, 28 Jun 2023 00:00:00 +0000

This research paper proposes a novel method, Exponential Moving Average Long Short-Term Memory (EMA LSTM), for multi-step vector output prediction of time series data using deep learning. The method combines the LSTM with the exponential moving average (EMA) technique to reduce noise in the data and improve the accuracy of prediction. The research compares the performance of EMA LSTM to other commonly used deep learning models, including LSTM, GRU, RNN, and CNN, and evaluates the results using statistical tests. The dataset used in this study contains daily stock market prices for several years, with inputs of 60, 90, and 120 previous days, and predictions for the next 20 and 30 days. The results show that the EMA LSTM method outperforms other models in terms of accuracy, with lower RMSE and MAPE values. This study has important implications for real-world applications, such as stock market forecasting and climate prediction, and highlights the importance of careful preprocessing of the data to improve the performance of deep learning models.

Texture Analysis of Citrus Leaf Images Using BEMD for Huanglongbing Disease Diagnosis

Wed, 28 Jun 2023 00:00:00 +0000

Plant diseases significantly threaten agricultural productivity, necessitating accurate identification and classification of plant lesions for improved crop quality. Citrus plants, belonging to the Rutaceae family, are highly susceptible to diseases such as citrus canker, black spot, and the devastating Huanglongbing (HLB) disease. Traditional approaches for disease detection rely on expert knowledge and time-consuming laboratory tests, which hinder rapid and effective disease management. Therefore, this study explores an alternative method that combines the Bidimensional Empirical Mode Decomposition (BEMD) algorithm for texture feature extraction and Support Vector Machine (SVM) classification to improve HLB diagnosis. The BEMD algorithm decomposes citrus leaf images into Intrinsic Mode Functions (IMFs) and a residue component. Classification experiments were conducted using SVM on the IMFs and residue features. The results of the classification experiments demonstrate the effectiveness of the proposed method. The achieved classification accuracies, ranging from 61% to 77% for different numbers of classes, the results show that the residue component achieved the highest classification accuracy, outperforming the IMF features. The combination of the BEMD algorithm and SVM classification presents a promising approach for accurate HLB diagnosis, surpassing the performance of previous studies that utilized GLCM-SVM techniques. This research contributes to developing efficient and reliable methods for early detection and classification of HLB-infected plants, essential for effective disease management and maintaining agricultural productivity.

Digital Image Processing Using YCbCr Colour Space and Neuro Fuzzy to Identify Pornography

Beki Subaeki, Yana Aditia Gerhana, Meta Barokatul Karomah Rusyana, Khaerul Manaf — Wed, 28 Jun 2023 00:00:00 +0000

Pornography is a severe problem in Indonesia, apart from drugs. This can be seen based on data from the Ministry of Communication and Informatics in 2021 which found 1.1 million pornographic content online. The increasing number of access to pornographic content sites on the internet can prove this. Several studies have been conducted to produce preventive formulas. However, this research flow has not been effective in solving the problem. This is because the results of the identification value in the output image obtained are not quite right. This study proposes a procedure for identifying pornographic content in digital images as an alternative approach for the early stages of a destructive content access prevention system. The formulation uses the YCbCr color space to analyze human skin on image objects that represent exposed body parts and the classification process with the Neuro Fuzzy approach. The performance of this formula was tested on 100 digital images of random categories of human objects (usually covered, skimpy, and naked) taken from the internet. The test results are at a relatively good level of accuracy, with a weight of 70% for the entire test data.