Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi)
Not a member yet
1071 research outputs found
Sort by
Classification of Secondary School Destination for Inclusive Students using Decision Tree Algorithm
Get PDFInclusive student education has become one of the most important agendas of UNESCO and the Indonesian government. Developing an inclusive education for children is critical to adapt to their abilities while attending school. However, most parents and educators who help students select their future secondary school after finishing primary school are often unaware of their real potential. The problem is mainly because the decision is not based on objective assessments such as IQ, average, and mental scores. In this study, our objective is to create a school-type decision support system using data mining as a factor-analytic approach to extract rules for the knowledge model. The system uses some variables as the basic principles for building school-type classification rules using the ID3 decision tree method. This system can also assist educators in making decisions based on existing graduate data. The evaluation showed that the proposed system produced an accuracy of 90% by allocating 75% of the data for training and 25% for testing. The accuracy value from the evaluation phase stated that the ID3 decision tree algorithm performs well. This system can also dynamically create new decision trees based on newly added datasets. More research is expected to result in a more variable and dynamic system that can have a more accurate result for the inclusive student classification of secondary school
Combination of K-NN and PCA Algorithms on Image Classification of Fish Species
Get PDFTo do fish farming, you need to know the types of fish to be cultivated. This is because the type of fish will affect how it is handled and managed. Therefore, this study aims to develop an image processing system for classifying fish species, especially cultivated fish, with a combination of the K-Nearest Neighbor (K-NN) algorithm and Principal Component Analysis (PCA). The feature extraction used is feature extraction based on its color and shape. The K-NN algorithm can group certain objects considering the shortest distance from the object. According to the best criteria, the PCA method is employed in the meanwhile to decrease and keep the majority of the relevant data from the original characteristics. On the basis of the test results, the accuracy value obtained is 85%. The use of a combination of the K-NN and PCA algorithms in the image classification of fish species in the research that has been done has been shown to be capable of increasing accuracy by 7.5% compared to only using the K-NN algorithm
Classification of Acute Lymphoblastic Leukemia based on White Blood Cell Images using InceptionV3 Model
Get PDFAcute lymphoblastic leukemia (ALL) is the most common form of leukemia that occurs in children. Detection of ALL through white blood cell image analysis can help with the prognosis and appropriate treatment. In this study, the author proposes an approach to classifying ALL based on white blood cell images using a convolutional neural network (CNN) model called InceptionV3. The dataset used in this research consists of white blood cell images collected from patients with ALL and healthy individuals. These images were obtained from The Cancer Imaging Archive (TCIA), which is a service that stores large-scale cancer medical images available to the public. During the evaluation phase, the author used training data evaluation metrics such as accuracy and loss to measure the model's performance. The research results show that the InceptionV3 model is capable of classifying white blood cell images with a high level of accuracy. This model achieves an average ALL recognition accuracy of 0.9896 with a loss of 0.031. The use of CNN models such as InceptionV3 in medical image analysis has the potential to improve the efficiency and precision of image-based disease diagnosis.Acute Lymphoblastic Leukemia (ALL) is the most prevalent form of leukemia that occurs in children. Detection of ALL through white blood cell image analysis can assist in prognosis and appropriate treatment. In this study, the author proposes an approach for classifying ALL based on white blood cell images using a Convolutional Neural Network (CNN) model called InceptionV3. The dataset used in this research consists of white blood cell images collected from patients with ALL and healthy individuals. These images were obtained from The Cancer Imaging Archive (TCIA), which is a service for storing large-scale cancer medical images available to the public. During the evaluation phase, the author used training data evaluation metrics such as accuracy and loss to measure the model's performance. The research results show that the InceptionV3 model is capable of classifying white blood cell images with a high level of accuracy. This model achieves an average ALL recognition accuracy of 0.9896 with a loss of 0.031. The use of CNN models like InceptionV3 in medical image analysis has the potential to enhance the efficiency and accuracy of image-based disease diagnosis
Leaf Image Identification: CNN with EfficientNet-B0 and ResNet-50 Used to Classified Corn Disease
Get PDFCorn is the second largest commodity in Indonesia after rice. In Indonesia, East Java is the largest corn producer. The first symptom of the disease in corn plants is marked by small brownish oval spots which are usually caused by the fungus Helminthoporium maydis, if left unchecked, farmers can suffer losses due to crop failure. Therefore it is important to provide treatment for diseases in corn plants as early as possible so that diseases in corn plants do not spread to other plants. In this study, the dataset used was taken from the kaggle website entitled Corn or Maize Leaf Disease Dataset. This dataset has 4 classifications: Blight, Common Rust, Grey leaf spot, and Healthy. This study uses the Convolutional Neural Network method with 2 different models, namely the EfficientNet-B0 and ResNet-50 models. The architectures used are the dense layer, the dropout layer, and the GlobalAveragePooling layer with a dataset sharing ratio of 70% which is training data and 30% is validation data. After testing the two proposed scenarios, the accuracy results obtained in the test model scenario 1, namely EfficientNet- B0 is 94% and for the second test model scenario, namely ResNet-50, the accuracy is 93%.Corn is the second largest commodity in Indonesia after rice. In Indonesia, East Java is the largest corn producer. The first symptom of the disease in corn plants is marked by small brownish oval spots which are usually caused by the fungus Helminthoporium maydis, if left unchecked, farmers can suffer losses due to crop failure. Therefore it is important to provide treatment for diseases in corn plants as early as possible so that diseases in corn plants do not spread to other plants. In this study, the dataset used was taken from the kaggle website entitled Corn or Maize Leaf Disease Dataset. This dataset has 4 classifications: Blight, Common Rust, Grey leaf spot, and Healthy. This study uses the Convolutional Neural Network method with 2 different models, namely the EfficientNet-B0 and ResNet-50 models. The architectures used are the dense layer, the dropout layer, and the GlobalAveragePooling layer with a dataset sharing ratio of 70% which is training data and 30% is validation data. After testing the two proposed scenarios, the accuracy results obtained in the test model scenario 1, namely EfficientNet- B0 is 94% and for the second test model scenario, namely ResNet-50, the accuracy is 93%
Folk Games Image Captioning using Object Attention
Get PDFThe result of a deep learning-based image captioning system with encoder-decoder framework relies heavily on the image feature extraction technique and the caption-based model. The accuracy of the model is heavily influenced by the proposed attention mechanism. The inability to distinguish between the output of the attention model and the input expectation of the decoder can cause the decoder to give incorrect results. In this paper, we proposed an object-attention mechanism using object detection. Object detection outputs a bounding box and an object category label, which is then used as an image input into VGG16 for feature extraction and into a caption-based LSTM model. The experimental results showed that the system with object attention performed better than the system without object attention. BLEU-1, BLEU-2, BLEU-3, BLEU-4, and CIDER scores for the image captioning system with object attention improved 12.48%, 17.39%, 24.06%, 36.37%, and 43.50% respectively compared to the system without object attention.
The result of deep-learning based image captioning system with encoder-decoder framework relies heavily on image feature extraction technique and caption-based model. The model accuracy is heavily influenced by the proposed attention mechanism. Unsuitability between the output of the attention model and the input expectation of the decoder can cause the decoder to give incorrect results. In this paper, we proposed an object attention mechanism using object detection. Object detection outputs a bounding box and object category label, which is then used as an image input into VGG16 for feature extraction and into a caption-based LSTM model. Experiment results showed that the system with object attention gave better performances than the system without object attention. BLEU-1, BLEU-2, BLEU-3, BLEU-4, and CIDER scores for image captioning system with object attention improved 12.48%, 17.39%, 24.06%, 36.37%, and 43.50% respectively compared to the system without object attention
Fire Detection System At Labuhanbatu University Based On Internet Of Things (IoT)
Get PDFFire accidents are disasters that often occur compared to other fire disasters such as floods, landslides, earthquakes or tsunamis. Fires can occur at any time, and no one knows for sure when a fire accident will occur. The impact of a fire disaster is not only material that can disappear from human lives. The causative factors of fire disasters often occur due to human negligence and fires often occur in houses where the occupants have left them. Labuhanbatu University at night will be left by the owner and all lecturers and educational staff, only guarded by two security people with this condition, it is very dangerous when a fire occurs in one of the buildings. The purpose of this research is to focus on developing a fire detection system at Labuhanbatu University based on the Internet of Things to provide early warning of safety. The system uses three sensors, namely temperature sensor, gas sensor, and fire sensor. This research is R&D research using the ADDIE model with the following stages: analysis, design, development, implementation, and evaluation. The results of the fire sensor test were 90% successful, the results of the sensor test as soon as possible were 90% successful, and the results of the temperature sensor test were 90% successful. This fire detection system can minimize or minimize the occurrence of fire accidents and losses because it is based on the Internet of Things providing early information when a fire occurs to education staff and lecturers at Labuhanabtu University. Overall, this fire warning system can function properly.
Fire accidents are disasters that often occur compared to other fire disasters such as floods, landslides, earthquakes or tsunamis. Fires can occur at any time and no one knows for sure when a fire accident will occur. The impact of a fire disaster is not only material that can disappear human lives. The causative factors of fire disasters often occur due to human negligence and fires often occur in houses where the occupants have left them. Labuhanbatu University at night will be left by the owner and all lecturers and educational staff, only guarded by 2 security people with this condition it is very dangerous when a fire occurs in one of the buildings. The purpose of this research is to focus on making a fire detection system at Labuhanbatu University based on the internet of things to provide early warning on safety. The system uses three sensors namely temperature sensor, gas sensor, and fire sensor. This research is an R&D research using the ADDIE model with the following stages Analysis, Design, Development, Implementation, Evaluation. The results of the fire sensor test were 90% successful, the results of the sensor test as soon as possible were 90% successful and the temperature sensor test results were 90% successful. This fire detection system can minimize or minimize the occurrence of fire accidents and losses because it is based on the internet of things providing early information when a fire occurs to education staff and lecturers at Labuhanabtu University. Overall this fire warning system can function properly
Systematic Mapping Study: Research Opportunities on Capacity Planning
Get PDFThe central idea of the research is to improve the efficiency and sustainability of data centers by implementing accurate capacity planning, which will also improve their performance and availability. Various literature reviews have been conducted to understand the current status of capacity planning implementation in different domains and perspectives. However, a more organized and systematic approach is required to map research and implementation results in the relevant areas of capacity planning that have the potential for further development. The present study aims to fill this gap by conducting a systematic mapping study that combines both quantitative and qualitative methodologies. The quantitative approach involved the collection of literature and the classification of topics using the Latent Dirichlet Allocation (LDA) method. On the contrary, the qualitative approach used content analysis to identify future research directions based on keyword trends and topics. The PRISMA framework was followed to guide the search for relevant studies in electronic research literature databases. The mapping results revealed 15 topics, with topics 8, 10, 11, and 15 showing significant potential for further research and exhibiting increasing trends. The identified topics encompass capacity planning, energy and resource management, computing and technology, data analysis and statistics, engineering, and industry, all crucial for businesses and industries to operate efficiently and sustainably. This study provides a comprehensive overview of the state of capacity planning implementation and highlights areas that require further investigation.
The central idea of the research is to enhance the efficiency and sustainability of data centers by implementing accurate capacity planning, which will also improve their performance and availability. Various literature reviews have been conducted to understand the current status of capacity planning implementation across different domains and perspectives. However, a more organized and systematic approach is required to map research and implementation outcomes in the relevant areas of capacity planning that have the potential for further development. The present study is aimed at filling this gap by conducting a systematic mapping study, which combines both quantitative and qualitative methodologies. The quantitative approach involved the collection of literature and topic classification using the Latent Dirichlet Allocation (LDA) method. In contrast, the qualitative approach utilized content analysis to identify future research directions based on keyword trends and topics. The PRISMA framework was followed to guide the search for relevant studies in electronic research literature databases. The mapping results revealed 15 topics, with topics 8, 10, 11, and 15 showing significant potential for further research and exhibiting increasing trends. The identified topics encompass capacity planning, energy and resource management, computing and technology, data analysis and statistics, engineering, and industry, all crucial for businesses and industries to operate efficiently and sustainably. This study provides a comprehensive overview of the state of capacity planning implementation and highlights areas that require further investigation
Precision Marketing Model using Decision Tree on e-Commerce Case Study Orebae.com
Get PDFThe development of the industrial world towards industry 4.0 has resulted in changes in the lifestyle of the wider community in carrying out their activities through digital media, one of which is shopping. This has an impact on the emergence of many business actors in the e-Commerce field, which brings its own challenges to stay alive and face the competition. The demands for innovation in competitive competition are also increasingly diverse with various approaches ranging from technology, social science, management science, and even artificial intelligence. One form of innovation that is widely carried out by e-Commerce today is looking for an ideal and effective form of marketing, where the form of marketing itself is considered less able to accommodate e-Commerce needs. One form of real innovation in finding the ideal and effective marketing is precision marketing. Precision marketing itself is marketing that is carried out by utilizing data where consumers are the center of preference for data collection. In fact, many of the e-commerce companies that were launched were unable to keep up with the competition because they were unable to develop marketing strategies and eventually went bankrupt. Therefore, we need a special way to bridge these problems so that e-Commerce can stay alive, especially for e-Commerce classified as Small and Medium Enterprises (SMEs). This research will focus on developing a precision marketing model in e-Commerce for small businesses, namely orebae.com which can be used as a tool in the development of marketing strategies. This research was carried out using a machine learning approach by adopting a decision tree algorithm. The results of this study showed that the precision marketing model for orebae.com based on customer preferences can be used to increase the number of sales of orebae.com and to reduce marketing costs.Perkembangan dunia industri menuju industri 4.0 semakin luas dalam perubahan gaya hidup masyarakat dalam menjalankan aktivitasnya melalui media digital salah satunya adalah berbelanja. Hal ini berdampak pada munculnya banyak pelaku usaha di bidang e-commerce yang membawa tantangan tersendiri untuk tetap hidup dan menghadapi persaingan. Tuntutan inovasi dalam persaingan yang kompetitif juga semakin beragam dengan berbagai pendekatan mulai dari teknologi, ilmu sosial, ilmu manajemen bahkan kecerdasan buatan. Salah satu bentuk inovasi yang banyak dilakukan oleh e-commerce saat ini adalah mencari bentuk pemasaran yang ideal dan efektif, dimana bentuk pemasaran itu sendiri dinilai kurang mampu mengakomodir kebutuhan e-commerce. Salah satu bentuk inovasi nyata dalam menemukan pemasaran yang ideal dan efektif adalah dengan pemasaran yang presisi. Pemasaran presisi itu sendiri adalah pemasaran yang dilakukan dengan memanfaatkan data dimana konsumen menjadi pusat preferensi untuk pengumpulan data. Bahkan, banyak perusahaan e-commerce yang diluncurkan tidak mampu mengikuti persaingan karena tidak mampu mengembangkan strategi pemasaran dan akhirnya bangkrut. Oleh karena itu, diperlukan suatu cara khusus untuk menjembatani permasalahan tersebut agar e-commerce dapat tetap hidup, khususnya bagi e-commerce yang tergolong Usaha Kecil Menengah (UKM). Penelitian ini akan fokus pada pengembangan model presisi pemasaran di e-commerce UKM yaitu orebae.com yang dapat digunakan sebagai alat dalam mengembangkan strategi pemasaran. Dari hasil penelitian ini diharapkan dapat menghasilkan model pemasaran orebae.com yang tepat sesuai dengan preferensi pelanggan, diharapkan dapat mempertahankan atau menambah jumlah konsumen orebae.com dan menekan biaya pemasaran. Untuk mencapai tujuan tersebut, penelitian ini dilakukan dengan menggunakan pendekatan machine learning dengan mengadopsi algoritma Decision Tree
Designing A WSNs-based Smart Home Monitoring System through Deep Reinforcement Learning
Get PDFThe technology of smart home systems has developed rapidly and provides convenience for human life. Several smart home technologies, especially monitoring systems, have been developed by integrating several aspects, including security systems, fuzzy methods, and energy saving methods. However, the problem is how to build a smart home system that is accurate, convenient, and low-cost. In this research, the development of a smart home monitoring system that integrates wireless sensor networks (WSNs) and deep reinforcement learning (DRL) is carried out based on three parameters, i.e. temperature, humidity and CO2 level. The experimental method is carried out by (1) validating the accuracy quality of WSNs; (2) determining the best model implemented in the system; and (3) measuring the quality of the DRL system on the smart home monitoring system. Based on the test results, several indicators were obtained: (1) WSN testing resulted in an accuracy of 98.52%; (2) the accuracy of the modeling results implemented in the system is 97.70%; and (3) DRL system test on the smart home monitoring system through 21 test scenarios resulted in an accuracy of 95.52%. The indicators of testing this smart monitoring system prove that the developed system provides the advantages of accuracy, ease of use, and low cost.
 
Implementation of Word Recommendation System Using Hybrid Method for Speed Typing Website
Get PDFTyping is one of the most frequently done activities in society therefore a medium is necessary to help train typing words that are often mistyped. Methods used in this research are the Content-Based Filtering Algorithm to gather the words that have a similar pattern to the words that are often mistyped based on the user's previous typing records and the Collaborative Filtering Algorithm that uses other users typing pattern to recommend the words. The result of this study shows the Collaborative Filtering Algorithm was able to gather words that are hard to type by the user with an accuracy of 49.2%, dan the Collaborative Filtering able to predict the score on how difficult for the user to type a word with the result of Root Mean Square Error (RMSE) value of 0.82 and with the Root Mean Square Percentage Error (RMSPE) value of 30% from the actual value, and a website which is the combination of the two algorithms with the result of 28% of the total word that is recommended was indeed difficult to type by the user with the typing speed of 103 WPM, and 72.3% for the user that has a typing speed of 39 WPM.Typing is one of the most frequently done activities in society therefore a medium is necessary to help train typing words that are often mistyped. Methods used in this research are the Content-Based Filtering Algorithm to gather the words that have a similar pattern to the words that are often mistyped based on the user's previous typing records and the Collaborative Filtering Algorithm that uses other users typing pattern to recommend the words. The result of this study shows the Collaborative Filtering Algorithm was able to gather words that are hard to type by the user with an accuracy of 49.2%, dan the Collaborative Filtering able to predict the score on how difficult for the user to type a word with the result of Root Mean Square Error (RMSE) value of 0.82 and with the Root Mean Square Percentage Error (RMSPE) value of 30% from the actual value, and a website which is the combination of the two algorithms with the result of 28% of the total word that is recommended was indeed difficult to type by the user with the typing speed of 103 WPM, and 72.3% for the user that has a typing speed of 39 WPM