Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi)
Not a member yet
1071 research outputs found
Sort by
Surrogate Model-based Multi-Objective Optimization in Early Stages of Ship Design
Get PDFThe abstract is the early stages of ship design, the decision of the ship's main dimensions significantly impacts the ship's performance and the total cost of ownership. This paper focuses on an optimization approach based on surrogate models at the early stages of ship design. The objectives are to minimize power requirements and building costs while still satisfying the constraints. We compare three approaches of surrogate models: Kriging, BPNN-PSO (Backpropagation Neural Network-Particle Swarm Optimizer), and MLP (Multi-Layer Perceptron) in two multi-objective optimization algorithms: MOEA/D (Multi-Objective Evolutionary Algorithm Decomposition) and NSGA-II (Non-Dominated Sorting Genetic Algorithm II). The experimental results show that MLP surrogate models get the best performance with MAE 6.03, and BPNN-PSO gets the second position with MAE 7.2. BPNN-PSO and MLP with MOEA/D and NSGA-II improve the design with around 58% smaller adequate power and 6% less steel weight than the original design. However, BPNN-PSO and MLP have lower hypervolume than Kriging for both optimization algorithms MOEA/D and NSGA-II. On the other hand, Kriging has the most inadequate model accuracy performance, with an MAE of 22.2, but produces the highest hypervolume, lowest computational time, and far lower objective values than BPNN-PSO and MLP for both optimization algorithms, MOEA/D and NSGA-II. Nevertheless, the three surrogate model approaches can significantly improve ship design solutions and reduce work time in the early stages of design.The abstract isthe early stages of ship design, the decision of the ship's main dimensions significantly impacts the ship's performance and the total cost of ownership. This paper focuses on an optimization approach based on surrogate models at the early stages of ship design. The objectives are to minimize power requirements and building costs while still satisfying the constraints. We compare three approaches of surrogate models: Kriging, BPNN-PSO (Backpropagation Neural Network-Particle Swarm Optimizer), and MLP (Multi-Layer Perceptron) in two multi-objective optimization algorithms: MOEA/D (Multi-Objective Evolutionary Algorithm Decomposition) and NSGA-II (Non-Dominated Sorting Genetic Algorithm II). The experimental results show that MLP surrogate models get the best performance with MAE 6.03, and BPNN-PSO gets the second position with MAE 7.2. BPNN-PSO and MLP with MOEA/D and NSGA-II improve the design with around 58% smaller adequate power and 6% less steel weight than the original design. However, BPNN-PSO and MLP have lower hypervolume than Kriging for both optimization algorithms MOEA/D and NSGA-II. On the other hand, Kriging has the most inadequate model accuracy performance, with an MAE of 22.2, but produces the highest hypervolume, lowest computational time, and far lower objective values than BPNN-PSO and MLP for both optimization algorithms, MOEA/D and NSGA-II. Nevertheless, the three surrogate model approaches can significantly improve ship design solutions and reduce work time in the early stages of design
Malaria Blood Cell Image Classification using Transfer Learning with Fine-Tune ResNet50 and Data Augmentation
No full textBased on the WHO Report related to malaria, it is estimated that there will be 241 million malaria cases and 627,000 deaths from this disease globally in 2020 with the number of deaths increasing yearly. Preventing malaria disease conditions is through early detection. A more quick and precise malaria diagnosis method was required to simplify and reduce the detection process. Medical image classification could be carried out rapidly and precisely using machine learning or deep learning techniques. This research aims to diagnose malaria by classifying images of malaria blood cells using Deep Learning with a Transfer Learning approach. By utilizing various fine-tuning procedures and implementing data augmentation proposed method develops the method from previous studies. Two types of models Frozen ResNet50 and Fine-Tune ResNet50 are being tested. The dataset utilized will be augmented to improve model performance. This study makes use of the "NIH Malaria Cell Images Dataset" a dataset that contains a total of 27,660 image data. It is divided into two classes: parasitized and uninfected. The results are improved from previous research using the fine-tuned VGG16 model with an accuracy of 96% compared to this study using the fine-tuned ResNet50 model which achieved an accuracy score of 98%.Based on the WHO Report related to malaria, it is estimated that there will be 241 million malaria cases and 627,000 deaths from this disease globally in 2020 with the number of deaths increasing yearly. Preventing malaria disease conditions is through early detection. A more quick and precise malaria diagnosis method was required to simplify and reduce the detection process. Medical image classification could be carried out rapidly and precisely using machine learning or deep learning techniques. This research aims to diagnose malaria by classifying images of malaria blood cells using Deep Learning with a Transfer Learning approach. By utilizing various fine-tuning procedures and implementing data augmentation proposed method develops the method from previous studies. Two types of models Frozen ResNet50 and Fine-Tune ResNet50 are being tested. The dataset utilized will be augmented to improve model performance. This study makes use of the "NIH Malaria Cell Images Dataset" a dataset that contains a total of 27,660 image data. It is divided into two classes: parasitized and uninfected. The results are improved from previous research using the fine-tuned VGG16 model with an accuracy of 96% compared to this study using the fine-tuned ResNet50 model which achieved an accuracy score of 98%
Gift Recommendations Based on Personality Using Fuzzy and Big Five Personality Test
No full textGifts are usually given to someone to strengthen a relationship or to motivate someone. However, givers often need help determining the appropriate gift for the potential recipient. On the other hand, many recipients are disappointed with the gifts received. This event can result in the relationship between the giver and recipient being disrupted or the motivational goal not being achieved. This research aims to develop a system to recommend gifts based on the recipient's personality. Gift recommendation is determined based on the recipient's personality because the recipient highly values gifts that match the recipient's personality. The system is built using the Fuzzy method, and the personality measurement tool used is the Big Five Personality Test. Fifteen pairs of respondents validated the system. The validation results show that 80% of respondents as gift-givers strongly agree that the system helps determine the appropriate gift for someone. In addition, 73.33% of respondents as gift recipients strongly agree that the gifts recommended by the system do not disappoint them.Gifts are usually given to someone to strengthen a relationship or to motivate someone. However, givers often need help determining the appropriate gift for the potential recipient. On the other hand, many recipients are disappointed with the gifts received. This event can result in the relationship between the giver and recipient being disrupted or the motivational goal not being achieved. This research aims to develop a system to recommend gifts based on the recipient's personality. Gift recommendation is determined based on the recipient's personality because the recipient highly values gifts that match the recipient's personality. The system is built using the Fuzzy method, and the personality measurement tool used is the Big Five Personality Test. Fifteen pairs of respondents validated the system. The validation results show that 80% of respondents as gift-givers strongly agree that the system helps determine the appropriate gift for someone. In addition, 73.33% of respondents as gift recipients strongly agree that the gifts recommended by the system do not disappoint them
Analisis Spasial Untuk Klasifikasi Pengembangan Tempat Penampungan Sementara Menggunakan Metode Jaringan Syaraf Tiruan
No full textGarbage is a problem that needs an in-depth study in urban areas because the development of an area has consequences on increasing population density, facilities and infrastructure, public services, and other aspects that impact increasing the volume of waste. The distribution of temporary waste shelters (TPS) in each area is still insufficient to accommodate the volume of waste, and its availability is inadequate. The purpose of this study is to model spatial data through spatial analysis using artificial intelligence methods in classifying the development of integrated temporary shelter locations (TPST) and regional integrated temporary shelters (TPST Regions) by utilizing Web-based technology (Geographical Information System (Web-GIS). The Artificial Neural Network method with the Backpropagation algorithm is used for the spatial analysis process based on the parameters of the population, the amount of organic and inorganic waste, the amount of industrial waste, and the volume of the TPST and Regional TPST capacity. The spatial analysis results using the Artificial Neural Network method obtained an accuracy value of 7171.02%. The results of this study can be the basis for Department of Environment and Cleanliness policies for the development of TPST and TPST areas with information coverage at the village level.Permasalahan sampah memerlukan perhatian serius akhir-akhir ini. Perkembangan daerah seiring dengan meningkatknya kepadatan penduduk, sarana dan prasarana, pelayanan publik, dan aspek lainnya serta berdampak pula pada peningkatan volume sampah. Persebaran tempat penampungan sementara (TPS) sampah yang disediakan di setiap daerah tidak sebanding dengan pertambahan volume sampah. Tujuan penelitian ini adalah untuk melakukan pemodelan data spasial melalui analisa spasial menggunakan metode kecerdasan buatan dalam mengklasifikasikan pengembangan lokasi penampungan sementara terpadu (TPST) dan tempat penampungan sementara terpadu kawasan (TPST Kawasan) dengan memanfaatkan teknologi SIG (Sistem Informasi Geografis) berbasis Web (Web-GIS). Metode Jaringan Syaraf Tiruan dengan algoritma Backpropagation digunakan dalam proses analisa spasial berdasarkan parameter jumlah penduduk, jumlah sampah penduduk organik dan an-organik, jumlah sampah industri, serta volume daya tampung TPST dan TPST Kawasan. Hasil analisa spasial dengan metode Jaringan Syaraf Tiruan memperoleh nilai akurasi 7171,02 %. Hasil penelitian ini dapat menjadi dasar kebijakan Dinas Lingkungan Hidup dan Kebersihan (DLHK) untuk pengembangan lokasi TPST dan TPST Kawasan dengan cakupan informasi pada level kelurahan atau desa
Pengenalan Emosi Pembicara Menggunakan Convolutional Neural Networks
Get PDFRecognition of the speaker's emotions is an important but challenging component of Human-Computer Interaction (HCI). The need for the recognition of the speaker's emotions is also increasing related to the need for digitizing the company's operational processes related to the implementation of industry 4.0. The use of Deep Learning methods is currently increasing, especially for processing unstructured data such as data from voice signals. This study tries to apply the Deep Learning method to classify the speaker's emotions using an open dataset from SAVEE which contains seven classes of voice emotions in English. The dataset will be trained using the CNN model. The final accuracy of the model is 88% on the training data and 52% on the test data, which means the model is overfitting. This is due to the imbalance of emotion classes in the dataset, which makes the model tend to predict classes with more labels. In addition, the lack of heterogeneity of the dataset makes the character of the emotion class more different from the others so that it can reduce the bias in the model so as not to overfit the model. Further development of this research can be done, such as over-sampling the existing dataset by adding other data sources, then performing data augmentation to get the data character of each emotion class and setting hyperparameter values โโto get better accuracy values.
Pengenalan emosi pembicara merupakan komponen penting namun menantang dari Interaksi Manusia-Komputer (HCI). Kebutuhan dari pengenalan emosi pembicara juga semakin meningkat terkait kebutuhan digitalisasi proses operasional perusahaan terkait implementasi industri 4.0. Penggunaan metode Deep Learning saat ini sudah semakin meningkat khususnya untuk pengolahan data tidak terstruktur seperti data dari sinyal suara. Penelitian ini mencoba mengimplementasikan metode Deep Learning untuk mengklasifikasikan emosi dari pembicara menggunakan dataset terbuka dari SAVEE yang berisi tujuh kelas emosi suara dalam bahasa Inggris. Dataset tersebut akan dilatih menggunakan model CNN. Akurasi akhir dari model sebesar 88% pada data latih dan 52% pada data tes yang berarti model mengalami overfitting. Hal tersebut disebabkan oleh tidak seimbangnya kelas emosi pada dataset sehingga membuat model akan cenderung memprediksi kelas yang labelnya lebih banyak. Selain itu, kurangnya heterogenitas dari dataset yang membuat karakter kelas emosi lebih berbeda dari yang lainnya sehingga mampu mengurangi bias pada model agar tidak membuat model overfitting. Pengembangan lebih lanjut dari penelitian ini yang dapat dilakukan seperti over sampling dataset yang dimiliki dengan menambahkan sumber data lain, melakukan augmentasi data agar mendapatkan karakter data setiap kelas emosi, dan mengatur nilai hyperparameter untuk mendapatkan nilai akurasi yang lebih baik
Knowledge Repository Design to Improve Knowledge Management Process Capabilities: A Systematic Literature Review
Get PDFThe role of technology in supporting knowledge management is very vital. The knowledge repository system is the foundation of knowledge management and its implementation. Knowledge sharing, discovery, and other knowledge processes will be more accessible through a proper knowledge repository system as an organizational knowledge base. This research explored various approaches employed by practitioners and researchers in designing a knowledge repository that allowed people to study and implement the knowledge repository as needed. A Systematic Literature Review (SLR) method proposed by Kitchenham was used to answer three research questions; (1) how is model knowledge storage in an organizational repository developed? (2) what tools can help create a knowledge repository? (3) What features should be present in a knowledge repository? The repository storage model was confirmed to be mainly represented in the ontology model. Furthermore, the technologies for creating knowledge were Protรฉgรฉ and Neo4j graph-based databases. In addition, features that were mostly applied in the knowledge repository system included data catalogs, search, API, knowledge management, visualization, rule-based and case-based reasoning.
 
The Generating Super Resolution of Thermal Image based on Deep Learning
Get PDFThe need for a high resolution to the thermal image is urgent and essential. The high resolution of the thermal image can give accurate information on the heat distribution map of the objects. The accurate heat distribution maps can give accurate temperature information. This accurate temperature measurement is used for measuring many objects such as electric motors, engines, the human body, and so onโthis information is used to detect the anomalies of the object to find the damaged parts. The anomalies are considered damaged parts found in solar panels, agricultural fields, buildings, bridges, etc. As the super-resolution of thermal images is very important, generating them is compulsory. The camera for obtaining super-resolution thermal images is rare, not available in the common market. Furthermore, this kind of device is costly too. Therefore not all the users, such as farmers or technicians, can have them. In order to handle the problem, the proposed method has the purpose of generating super-resolution thermal images economically and is more accessible through the deep learning method. The dataset is taken from the solar panel. The results show that the proposed method can handle the low-resolution problem of thermal images.
 
The Implementation of MQTT Protocol using PT-100 for Monitoring the Vaccine Temperature
Get PDFA vaccine has an essential role for humans because it can reduce the effect of organism infection and activate the body's immune. The limitation of temperature when the vaccine is stored becomes one of the influential factors for its quality. Since the recent technique used for checking the temperature is manual, the process of real-time temperature monitoring cannot be done and becomes the main problem in the process of storing vaccines. This research is aimed to develop a system for measuring temperature on a chiller through a temperature sensor and the internet of things automatically. The methods used in the MQTT protocol are published and subscribed. The quality of service (QoS) to measure network capability and PT-100 is the sensor in this research because it has a high accuracy level, linear signal change, and fast response. This sensor can measure glycol liquid with a temperature range of 2-8ยฐC. Data distribution uses an ESP module to connect the data to Antares through the wifi network. The temperature value measurement using PT-100 toward vaccine chiller temperature obtains an accuracy value of 99,32%. This accuracy value shows that the result is compatible with the logger data measurement tool development used with an error result of 0.58%. Delay parameter testing on the quality of service level obtains the average value of 41.91 ms, while the Packet loss parameter obtains the value of 0%.Vaksin adalah komponen yang sangat penting bagi kehidupan manusia. Salah satu faktor yang mempengaruhi kualitas vaksin yaitu penyimpanannya, karena beberapa jenis vaksin memerlukan penyimpanan khusus agar kualitasnya terjaga. Kualitas sebuah vaksin dapat terjaga dengan menjaga suhu agar tetap pada batas ketentuannya. Proses penyimpanan vaksin biasanya ditempatkan pada chiller. Penyimpanan membutuhkan proses monitoring suhu secara real time, beberapa masalah diantaranya pengecekan yang masih terjadi kesalahan pencatatan sehingga tidak efektif dan efisien. Tujuan yang hendak dicapai dari penelitian ini adalah mengembangkan system monitoring suhu chiller vaksin menggunakan sensor PT-100 dengan protokol MQTT sehingga kegiatan monitoring dapat efektif, efisien dan meminimalisir kesalahan operator. Metode pada protokol MQTT terdapat dua sisi yaitu sebagai publish dan subscribe. Sensor PT-100 digunakan untuk mengetahui suhu pada cairan glikol yang digunakan untuk mewakili cairan vaksin yang baik direntang suhu 2-8ยฐC dan untuk mikrokontroler menggunakan modul ESP32 yang sudah terdapat WiFi untuk pengiriman datanya. Protokol komunikasi MQTT yang akan di hubungkan dengan platform Antares sebagai broker. Hasil pengujian sensor PT-100 terhadap suhu chiller vaksin mendapatkan nilai akurasi sebesar 99,32%. Dari hasil pengujian nilai pembacaan sensor sudah sesuai dengan data yang ada pada chiller ini menandakan system yang dirancang sudah bekerja dengan baik. Selain itu pengujian QoS (Quality of Service) didapatkan nilai rata-rata 41,91 ms. Pada pengujian parameter Packet loss mendapatkan nilai 0%. Semua hasil pengujian Delay dan Packet loss menurut standar TIPHON masuk dalam kriteria sangat bagus.  
The Formula Study in Determining the Best Number of Neurons in Neural Network Backpropagation Architecture with Three Hidden Layers
Get PDFThe researchers conducted data simulation experiments, but they did so unstructured in determining the number of neurons in the hidden layer in the Artificial Neural Network Back-Propagation architecture. The researchers also used a general architecture consisting of one hidden layer. Researchers are still producing minimal research that discusses how to determine the number of neurons when using hidden layers. This article examines the results of experiments by conducting training and testing data using seven recommended formulas including the Hecht-Nelson, Marchandani-Cao, Lawrence & Fredrickson, Berry-Linoff, Boger-Guterman, JingTao-Chew, and Lawrence & Fredrickson modifications. We use rainfall data and temperature data with a 10-day type for the last 10 years (2012-2021) sourced from Lombok International Airport Station, Indonesia. The training and testing data used showed the results that in determining the number of neurons on the hidden-1 screen, it was more appropriate to use the Hecht-Nelson formula and the Lawrence & Fredricson formula which is more suitable for use in the 2nd & 3rd hidden layer. The resulting research was able to provide an accuracy rate of up to 97.79% (temperature data) and 99.94% (rainfall data) with an architecture of 36-73-37-19-1.
The researchers conducted data simulation experiments, but they did so unstructured in determining the number of neurons in the hidden layer in the Artificial Neural Network Back-Propagation architecture. The researchers also used a general architecture consisting of one hidden layer. Researchers are still producing minimal research that discusses how to determine the number of neurons when using hidden layers. This article examines the results of experiments by conducting training and testing data using seven recommended formulas including the Hecht-Nelson, Marchandani-Cao, Lawrence & Fredrickson, Berry-Linoff, Boger-Guterman, JingTao-Chew, and Lawrence & Fredrickson modifications. We use rainfall data and temperature data with a 10-day type for the last 10 years (2012-2021) sourced from Lombok International Airport Station, Indonesia. The training and testing data used showed the results that in determining the number of neurons on the hidden-1 screen, it was more appropriate to use the Hecht-Nelson formula and the Lawrence & Fredricson formula which is more suitable for use in the 2nd & 3rd hidden layer. The resulting research was able to provide an accuracy rate of up to 97.79% (temperature data) and 99.94% (rainfall data) with an architecture of 36-73-37-19-1.
 
Fuzzy Learning Vector Quantization Untuk Klasifikasi Citra Daging Oplosan Berdasarkan Ciri Warna dan Tekstur
No full textBeef consumption is quite high and expensive in the world. In Indonesia, beef prices are relatively expensive because the meat supply chain from farmers to the market is quite long. The high demand for beef and the difficulty of obtaining meat are factors in the high price of meat. This makes some meat traders cheat by mixing beef and pork (oplosan). Mixing beef and pork is detrimental to beef consumers, especially those who are Muslim. In this paper, we proposed a new strategy for identifying beef, pig, and mixed meat utilizing Fuzzy learning vector quantization (FLVQ) Based on the color and texture aspects of the meat. The HSV (Hue saturation value) approach is used for color features, whereas the GLCM (Gray level co-occurrence matrix) method is used for texture features. This study makes use of primary data collected from the Pasar Bawah Tourism and Cipuan Market in Pekanbaru, Riau Province. The data set consists of 600 photos, 200 each of beef, pork, and mixed. Based on the test scenario, the coefficient of fuzzyness and learning rate affect the accuracy of meat image identification. The proposed strategy has succeeded in classifying pork, beef and mixed meat with the best percentage of accuracy results in theclasses of beef and pork, beef and mixed, pork and mixed meat, respectively, at 100%, 97.5%, and 95%. This demonstrates that the proposed strategy has succeeded in classifying the image of pork, beef, and mixed.
Beef consumption is quite high and expensive in the world. In Indonesia, beef prices are relatively expensive because the meat supply chain from farmers to the market is quite long. The high demand for beef and the difficulty of obtaining meat are factors in the high price of meat. This makes some meat traders cheat by mixing beef and pork (oplosan). Mixing beef and pork is detrimental to beef consumers, especially those who are Muslim. In this paper, we proposed a new strategy for identifying beef, pig, and mixed meat utilizing Fuzzy learning vector quantization (FLVQ) Based on the color and texture aspects of the meat. The HSV (Hue saturation value) approach is used for color features, whereas the GLCM (Gray level co-occurrence matrix) method is used for texture features. This study makes use of primary data collected from the Pasar Bawah Tourism and Cipuan Market in Pekanbaru, Riau Province. The data set consists of 600 photos, 200 each of beef, pork, and mixed. Based on the test scenario, the coefficient of fuzzyness and learning rate affect the accuracy of meat image identification. The proposed strategy has succeeded in classifying pork, beef and mixed meat with the best percentage of accuracy results in theclasses of beef and pork, beef and mixed, pork and mixed meat, respectively, at 100%, 97.5%, and 95%. This demonstrates that the proposed strategy has succeeded in classifying the image of pork, beef, and mixed