Proceeding of the Electrical Engineering Computer Science and Informatics
Not a member yet
649 research outputs found
Sort by
Implementation of Role-Based Access Control on OAuth 2.0 as Authentication and Authorization
Get PDFAs today’s technology transition from monolithic towards microservices architecture, the authentication and authorization system also becomes a new concern because of the difference between monolithic and microservices pattern. Monolithic mostly uses role-based access control while microservices uses scope with OAuth 2.0. With this in mind, there is a need for a model that can integrate OAuth 2.0 with role-based access control. With rolebased access control implemented on OAuth 2.0, we expect a simpler authorization process and a more secure authentication and authorization system for microservices backend architecture. This paper proposes a model to implement role-based access control on OAuth 2.0 using Laravel framework, we also test the performance of the system following by response time, data transferred and throughput. From the performance test, this approach has a good performance and can handle certain requests with simulated users even with limited resources
Paraphrase Detection Using Manhattan's Recurrent Neural Networks and Long Short-Term Memory
Get PDFNatural Language Processing (NLP) is a part of artificial intelligence that can extract sentence structures from natural language. Some discussions about NLP are widely used, such as Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM) to summarize papers with many sentences in them. Siamese Similarity is a term that applies repetitive twin network architecture to machine learning for sentence similarity. This architecture is also called Manhattan LSTM, which can be applied to the case of detecting paraphrase sentences. The paraphrase sentence must be recognized by machine learning first. Word2vec is used to convert sentences to vectors so they can be recognized in machine learning. This research has developed paraphrase sentence detection using Siamese Similarity with word2vec embedding. The experimental results showed that the amount of training data is dominant to the new data compared to the number of times and the variation in training data. Obtained data accuracy, 800,000 pairs provide accuracy reaching 99% of training data and 82.4% of new data. These results are better than the accuracy of the new data, with half of the training data only yielding 64%. While the amount of training data did not effect on training data
Optimized Fuzzy Backpropagation Neural Network using Genetic Algorithm for Predicting Indonesian Stock Exchange Composite Index
Get PDFInvestment activities in the capital market have the possibility to generate profits and at the same time also cause losses. The composite stock price index as an indicator used to determine investment continues to change over time. Uncertainty of stock exchange composite index requires investors to be able to make predictions so as to produce maximum profits. The aim of this study is to forecast the composite stock price index. The input variables used are Indonesia interest rates, rupiah exchange rates, Dow Jones index, and world gold prices. All data obtained in the period from January 2008 to March 2019. Data are used to build the Fuzzy Backpropagation Neural Network (FBPNN), model. The weight of FBPNN model was optimized using Genetic Algorithm then used to forecast the composite stock price index. The forecasting result of the composite stock price index for April to June 2019 respectively were 5822.6, 5826.8, and 5767.3 with the MAPE value of 8.42%. These results indicate that Indonesia interest rates, rupiah exchange rate, Dow Jones index, and the gold price are the proper indicators to predict the composite stock price index
Design and Implementation of MPC for Energy Optimization of Boiler in Batch Distillation Column
Get PDFA competency in the industrial world depends on several aspects, that is cost, delivery, flexibility, and quality. The smart industrial system / Smart Manufacturing System (SMS) tries to improve those aspects using the latest technology that encourages the use of digital information widely and quickly in industrial systems. Development of SM in industry 4.0, pushing the change of industrial pyramids into CyberPhysical System (CPS). CPS has begun to be applied widely in process industries nowadays, i.e., distillation process industry. In distillation process, boiler has the important role to separate 2 different components using the difference of its boiling point. In this paper, the alcohol distillation plant is used to purify 30% of alcohol solution. The modelling of boiler, simulation, and implementation of boiler control system are presented to get the desired temperature. The temperature reference is roughly 85 oC. Predictive Model Controller (MPC) and Kalman Filter is proposed to control the temperature of boiler by adjusting the PWM of on-off time and to deal with the disturbance and sensor noise. The IAE, ISE, and ITAE is analyzed to obtain the error of control system and energy usage per operation is also calculated to find out the effect of MPC controller in energy optimizatio
Comparison of PIC and SIC with Lattice Reduction cancellation schemes for V-BLAST MIMO system
Get PDFThis paper discusses the design of the MIMO signal demodulation algorithm built according to the V-BLAST scheme. Several algorithms such as the maximum likelihood algorithm and QR-M algorithm, Minimum of Mean Square Error (MMSE), Successive Interference Cancellation (SIC) with MMSE (SICMMSE), Parallel Interference Cancellation (PIC) with MMSE (PIC-MMSE), Lattice Reduction MMSE (LR-MMSE), as well as modifications of SIC-LR-MMSE, PIC-LR-MMSE are considered. It is shown that the SIC-LR-MMSE and PIC-LR-MMSE algorithms, presented in this paper, having a linear computational complexity depending on the number of antennas and modulation order, provides good performance close to potentially achievable (MLA), which makes its practical use possible. The presented algorithm can be used to improve performance of CDN (Content Delivery Networks) or CMS (Content Management Systems) to transmit content in different distributed content delivery systems
Analysis of User Readiness Level of E-Government Using Stope Framework
Get PDFThe Village Information Administration System (VIAS) is one of the many IT service implemented by the Bondowoso district. The successful implementation of an IT service like VIAS can be seen from the extent to which the level of readiness of the user adopts the service. To determine the extent to which the user readiness level can be performed a measurement using the relevant method. STOPE Framework is one of the method that can be used to measure the extent to which the IT service user readiness and can provide recommendations based on the result of the evaluation. Measurement of user readiness in this study using STOPE Framework and determination of a 4 point scale based readiness level issued by the Centre for International Development/CID (Harvard Cyber Law). Data analysis result show VIAS user readiness level is at level 4(completely ready). This indicates that the VIAS user is ready to use the Village Information Administration System based on approach using STOPE Framework
IDEnet : Inception-Based Deep Convolutional Neural Network for Crowd Counting Estimation
Get PDFIn crowd counting task, our goals are to estimate density map and count of people from the given crowd image. From our analysis, there are two major problems that need to be solved in the crowd counting task, which are scale invariant problem and inhomogeneous density problem. Many methods have been developed to tackle these problems by designing a dense aware model, scale adaptive model, etc. Our approach is derived from scale invariant problem and inhomogeneous density problem and we propose a dense aware inception based neural network in order to tackle both problems. We introduce our novel inception based crowd counting model called Inception Dense Estimator network (IDEnet). Our IDEnet is divided into 2 modules, which are Inception Dense Block (IDB) and Dense Evaluator Unit (DEU). Some variations of IDEnet are evaluated and analysed in order to find out the best model. We evaluate our best model on UCF50 and ShanghaiTech dataset. Our IDEnet outperforms the current state-of-the-art method in ShanghaiTech part B dataset. We conclude our work with 6 key conclusions based on our experiments and error analysis
Re-Ranking Image Retrieval on Multi Texton Co-Occurrence Descriptor Using K-Nearest Neighbor
Get PDFSome features commonly used to conduct image retrieval are color, texture and edge. Multi Texton Co-Occurrence Descriptor (MTCD) is a method which uses all three features to perform image retrieval. This method has a high precision when doing retrieval on a patterned image such as Batik images. However, for images focusing on object detection like corel images, its precision decreases. This study proposes the use of KNN method to improve the precision of MTCD method by re-ranking the retrieval results from MTCD. The results show that the method is able to increase the precision by 0.8% for Batik images and 9% for corel images
Automatic User-Video Metrics Creations From Emotion Detection
Get PDFIn this digital era, digital content especially video, is increasing in number from time to time. Typically, a video service provider like Youtube will perform video analysis based on the video content such as colours, textures, shapes, and other features that exist in video content. The result of this analysis was used to understand user preference and to personalize video for each user. With technological developments, especially in Machine Learning and Computer Vision technology, video analysis can be based on other things beyond the video. In this context, it is the audience's impression. Thus, with the analysis of audience impressions in real-time, it is expected that the video can be analysed using the emotion parameters of the audience while the video is playing, and this can be done automatically and real-time. This system generates impression statistic for each video which concluded from every user who has watched the video and save those data in the database. Method used to analyse the result is by recruiting respondent and give some questionnaires. Respondents were asked to watch some videos and were asked to compare the impression metric which created by the system with user's real impression. The result shos that the automatic video-metric creation from emotion detection has been able to measure user's impression of the video with more than 80% accuracy stated by 75% of 20 respondents of the survey
Real Time SIBI Sign Language Recognition Based on K-Nearest Neighbor
Get PDFPersons with disabilities also have the right to communicate with each other, both with normal people and people with other disabilities. People with disabilities will be difficult to communicate with other people. They use 'sign language' to communicate. That's why other normal people will be difficult to communicate with them. Because there are not many normal people that can understand the 'sign language'. System which can help to communicate with disabilities people are needed. In this paper, we proposed sign language recognition for Sistem Isyarat Bahasa Indonesia (SIBI) using leap motion based on K-Nearest Neighbor. Technology of leap motion controller will generate the existence of coordinate points on each bone in hand. As an input, we used the value of distance between the coordinates of each bone distal to the position of the palm, which were measured using Euclidean Distance. This feature of distance will be used for training and testing data on K-Nearest Neighbor method. The experiment result shows that the best accuracy is 0,78 and error 0,22 with proposed parameter of K = 5