Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi)
Not a member yet
1071 research outputs found
Sort by
Deep Learning Implementation using Convolutional Neural Network for Alzheimer’s Classification
Get PDFAlzheimer's disease is the most common cause of dementia. Dementia refers to brain symptoms such as memory loss, difficulty thinking and problem solving and even speaking. This stage of development of neuropsychiatric symptoms is usually examined using magnetic resonance images (MRI) of the brain. The detection of Alzheimer's disease from data such as MRI using machine learning has been the subject of research in recent years. This technology has facilitated the work of medical experts and accelerated the medical process. In this study we target the classification of Alzheimer's disease images using convolutional neural network (CNN) and transfer learning (VGG16 and VGG19). The objective of this study is to classify Alzheimer's disease images into four classes that are recognized by medical experts and the results of this study are several evaluation metrics. Through experiments conducted on the dataset, this research has proven that the algorithm used is able to classify MRI of Alzheimer's disease into four classes known to medical experts. The accuracy of the first CNN model is 75.01%, the second VGG16 model is 80.10% and the third VGG19 model is 80.28%.Alzheimer's disease is the most common cause of dementia. Dementia refers to brain symptoms such as memory loss, difficulty thinking and problem solving and even speaking. This stage of development of neuropsychiatric symptoms is usually examined using magnetic resonance images (MRI) of the brain. The detection of Alzheimer's disease from data such as MRI using machine learning has been the subject of research in recent years. This technology has facilitated the work of medical experts and accelerated the medical process. In this study we target the classification of Alzheimer's disease images using convolutional neural network (CNN) and transfer learning (VGG16 and VGG19). The objective of this study is to classify Alzheimer's disease images into four classes that are recognized by medical experts and the results of this study are several evaluation metrics. Through experiments conducted on the dataset, this research has proven that the algorithm used is able to classify MRI of Alzheimer's disease into four classes known to medical experts. The accuracy of the first CNN model is 75.01%, the second VGG16 model is 80.10% and the third VGG19 model is 80.28%.
 
Decentralized Finance (DeFi), Strengths Become Weaknesses: a Literature Survey
No full textThe use of blockchain technology in Decentralized Finance (DeFi) has gained popularity, with 23 public companies and one country holding bitcoin. DeFi aims to create an open and decentralized financial ecosystem that is accessible to everyone, eliminates intermediaries like financial institutions, and is verifiable, immutable, globally accepted, fast, low-cost, anonymous, and non-custodial. Despite its benefits, the rapid growth of DeFi has led to increased security risks. This study assesses the validity of DeFi's superiority claims in light of security incidents and events in 2022 and Twitter trends. This study used a Systematic Literature Review from various research articles and news from 2022. This research found that DeFi's superiority claims seem to be inconsistent with what is being advertised. It also found that if DeFi is not properly prepared and audited, its strength (Anonymous, open-source, decentralized, non-custodial, eliminates third parties and regulation) may become its weakness. Despite this, users still exhibit high levels of trust and optimism, as seen in the most popular terms shared by user tweets during significant losses, with 301,654 unique tweets between April 30 and May 31, 2022 and 344,519 unique tweets between October 3 and December 3, 2022, that are crypto, nft, and blockchain.The use of blockchain technology in Decentralized Finance (DeFi) has gained popularity, with 23 public companies and one country holding bitcoin. DeFi aims to create an open and decentralized financial ecosystem that is accessible to everyone, eliminates intermediaries like financial institutions, and is verifiable, immutable, globally accepted, fast, low-cost, anonymous, and non-custodial. Despite its benefits, the rapid growth of DeFi has led to increased security risks. This study assesses the validity of DeFi's superiority claims in light of security incidents and events in 2022 and Twitter trends. This study used a Systematic Literature Review from various research articles and news from 2022. This research found that DeFi's superiority claims seem to be inconsistent with what is being advertised. It also found that if DeFi is not properly prepared and audited, its strength (Anonymous, open-source, decentralized, non-custodial, eliminates third parties and regulation) may become its weakness. Despite this, users still exhibit high levels of trust and optimism, as seen in the most popular terms shared by user tweets during significant losses, with 301,654 unique tweets between April 30 and May 31, 2022 and 344,519 unique tweets between October 3 and December 3, 2022, that are crypto, nft, and blockchain
A Systematic Literature Review of Automation Quality of Service in Computer Networks: Research Trends, Datasets, and Methods
Get PDFThe article is a systematic literature review of the use of automation for quality of service (QoS) in computer networks. It summarizes the research trends, datasets, and methods used in the field and provides an overview of the current state of the art. The focus of the review is on the use of automation for QoS management and improvement in computer networks, including the use of machine learning, artificial intelligence, and other computational techniques. The review highlights the need for further research and development in this area and provides insights into future directions for the field. The review covered a wide range of studies, including research papers and conference proceedings, and involved a comprehensive database search of the Scopus database covering journals and proceedings such as the Institute of Electrical and Electronics Engineers (IEEE) Xplore, Association for Computing Machinery (ACM) Digital Library, Springer, and ScienceDirect databases between 2017 and September 2022. From these databases, 1856 metadata were found, which eventually became seventy-three metadata after going through the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol.The article is a systematic literature review of the use of automation for quality of service (QoS) in computer networks. It summarizes the research trends, datasets, and methods used in the field and provides an overview of the current state of the art. The focus of the review is on the use of automation for QoS management and improvement in computer networks, including the use of machine learning, artificial intelligence, and other computational techniques. The review highlights the need for further research and development in this area and provides insights into future directions for the field. The review covered a wide range of studies, including research papers and conference proceedings, and involved a comprehensive database search of the Scopus database covering journals and proceedings such as the Institute of Electrical and Electronics Engineers (IEEE) Xplore, Association for Computing Machinery (ACM) Digital Library, Springer, and ScienceDirect databases between 2017 and September 2022. From these databases, 1856 metadata were found, which eventually became seventy-three metadata after going through the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol
Diabetes Risk Prediction using Feature Importance Extreme Gradient Boosting (XGBoost)
No full textDiabetes results from impaired pancreatic function as a producer of insulin and glucagon hormones, which regulate glucose levels in the blood. People with diabetes today are not only experienced adults, but pre-diabetes has been identified since the age of children and adolescents. Early prediction of diabetes can make it easier for doctors and patients to intervene as soon as possible so that the risk of complications can be reduced. One of the uses of medical data from diabetes patients is to produce a model that medical personnel can use to predict and identify diabetes in patients. Various techniques are used to provide the earliest possible prediction of diabetes based on the symptoms experienced by diabetic patients, including the use of machine learning. People can use machine learning to generate models based on historical data from diabetic patients, and predictions are made with the model. In this study, extreme gradient boosting is the machine learning technique for predicting diabetes (xgboost) using XGBoost with importance features. The diabetes dataset used in this study comes from the early stage diabetes risk prediction dataset published by UCI Machine Learning, which has 520 records and 16 attributes. The diabetes prediction model using xgboost is displayed as a tree. The model precision result in this study was 98.71%, for the F1 score was 98.18%. The accuracy obtained based on the best 10 attributes using the importance of the XGBoost feature is 98.72%.
Diabetes results from impaired pancreas function as a producer of insulin and glucagon hormones, which regulate glucose levels in the blood. People with diabetes today are not only experienced adults, but pre-diabetes has been identified since the age of children and adolescents. Early prediction of diabetes can make it easier for doctors and patients to intervene as soon as possible so that the risk of complications can be reduced. One of the uses of medical data from diabetes patients is used to produce a model that can be used by medical staff to predict and identify diabetes in patients. Various techniques are used to provide the earliest possible prediction of diabetes based on the symptoms experienced by diabetic patients, including using machine learning. People can use Machine Learning to generate models based on historical data of diabetic patients, and predictions are made with the model. In this study, extreme gradient boosting is the machine learning technique to predict diabetes (xgboost) using Feature Importance XGBoost. The diabetes dataset used in this study comes from the Early stage diabetes risk prediction dataset published by UCI Machine Learning, which has 520 records and 16 attributes. The diabetes prediction model using xgboost is displayed as a tree. The model accuracy result in this study was 98.71%, for the F1 score was 98.18%. While the accuracy obtained based on the best 10 attributes using the XGBoost feature importance are 98.72%
Transforming LMS into KMS in Indonesia Educational Institution Case Study in Telkom University Open Library
Get PDFLibrary is an institution that collects printed and recorded knowledge, manages it in a particular way to meet the intellectual needs of its users through various ways of knowledge interaction. In 2021, Indonesia provided 4605 active educational institutions, ranging from universities to community academies. All these institutions are obliged to provide library to create a better learning environment, by providing source of references for each course delivered by the institution. This obligation is encouraged by the government by considering library support within the accreditation system. In this accreditation system, a library should allocate books as a source of reference to each course. This establishes a paradigm that library is where we store books and where a member of the institution may borrow, learn, and return those aforementioned books. Today, educational institution deals with not only books but also thesis, dissertation, technical report, training/workshop report, research paper, etc. Authors believe it will be prudent to leave all these documents of knowledge to librarians, by changing the library's paradigm from managing books to managing knowledge. This study proposes a model of Knowledge Management System as a transformation from Library Management System. This study also explains about expected opportunities and benefits after the transformation.
Library is an institution that collects printed and recorded knowledge, manages it in a particular way to meet the intellectual needs of its users through various ways of knowledge interaction. In 2021, Indonesia provided 4605 active educational institutions, ranging from universities to community academies. All these institutions are obliged to provide library to create a better learning environment, by providing source of references for each course delivered by the institution. This obligation is encouraged by the government by considering library support within the accreditation system. In this accreditation system, a library should allocate books as a source of reference to each course. This establishes a paradigm that library is where we store books and where a member of the institution may borrow, learn, and return those aforementioned books. Today, educational institution deals with not only books but also thesis, dissertation, technical report, training/workshop report, research paper, etc. Authors believe it will be prudent to leave all these documents of knowledge to librarians, by changing the library's paradigm from managing books to managing knowledge. This study proposes a model of Knowledge Management System as a transformation from Library Management System. This study also explains about expected opportunities and benefits after the transformation
Utilization of Mobile Network Infrastructure to Prevent Financial Mobile Application Account Takeover
Get PDFThe Covid-19 pandemic has kept almost everyone at home and forced them to perform online activities using their mobile gadgets. Penetration of the Internet and mobile use is increased as lockdowns or restrictions on meeting face to face are getting used to. This has become a new market for cyber criminals to carry out their actions, such as spreading Social Engineering, sending Phishing, doing Account Takeover, and ending in theft of money in Financial Mobile Applications. Application protection with OTP SMS and Magic Link SMS still has vulnerabilities, with several examples of cases that have occurred. For this reason, this problem was raised to find a solution using the Mobile Network Infrastructure. The method used is to compare the congruence between the phone numbers registered in the application and the phone numbers used. Every time a user signs in or signs up, the Financial Mobile Application will perform Mobile Network Verification to cellular operators via API. Verification is carried out by utilizing the header enrichment in the background of the application process that was installed on the user's smartphone or tablet to the Mobile Network Verification Server. The Financial Mobile Applications can then determine whether the user is using a valid or invalid telephone number. Therefore, the target account cannot be taken over because the cyber criminal's mobile device does not have the phone number attached to the victim’s mobile device. This proof is carried out with four test case scenarios with the sign-up and sign-in processes on the same phone number and different phone numbers between devices and applications. It is hoped that this kind of protection model can reduce losses experienced by users of Financial Mobile Applications due to Account Takeover.Pandemi Covid-19 membuat hampir semua orang harus berdiam diri di rumah dan memaksa mereka untuk beraktivitas secara daring menggunakan mobile gadget. Penetrasi penggunaan internet dan mobile semakin meningkat seiring pembiasaan lockdown atau pemberlakuan pembatasan kegiatan masyarakat. Hal ini menjadi pasar baru bagi para penjahat dunia maya untuk melakukan aksinya, seperti melakukan Social Engineering, mengirimkan Phishing, melakukan Account Take Over, hingga berujung pada pencurian uang di aplikasi seluler finansial. Proteksi aplikasi dengan SMS OTP dan SMS Magic Link masih memiliki kerentanan dengan beberapa contoh kasus yang pernah terjadi. Untuk itu permasalahan ini diangkat untuk dicarikan solusinya dengan memanfaatkan Mobile Network Infrastructure. Metode yang digunakan adalah membandingkan kesesuaian antara nomor telepon yang terdaftar di aplikasi dengan nomor telepon yang digunakan. Setiap kali pengguna masuk atau mendaftar, aplikasi seluler finansial akan melakukan Mobile Network Verification ke operator seluler melalui API. Verifikasi dilakukan dengan memanfaatkan header enrichment di latar belakang proses aplikasi yang diinstal pada smartphone atau tablet pengguna ke Mobile Network Verification Server. Kemudian aplikasi seluler finansial dapat mengetahui, pengguna menggunakan nomor telepon yang valid atau tidak valid. Dengan demikian, akun target tidak dapat diambil alih, karena penjahat dunia maya tidak memiliki nomor telepon korban yang terpasang di-gadget-nya. Pembuktian ini dilakukan dengan empat skenario test case dengan proses sign-up dan sign-in pada nomor telepon yang sama dan berbeda antar perangkat dan aplikasi. Model perlindungan seperti ini diharapkan dapat mengurangi kerugian yang dialami pengguna Financial Mobile Application akibat Account Take Over
Evaluating Player Experience for Fear Modeling of 2D East Java Horror Game Alas Tilas
Get PDFDeveloping a 2D horror game and evaluating the reliability of the player experience are two things that are interrelated and equally important. Developers must ensure that the game can provide a satisfying and reliable gaming experience for their players. This study aims to evaluate the reliability of the player's experience in the game entitled Alas Tilas, East Java. This study used the User Experience Questionnaire (UEQ) in Indonesian as a survey approach method, which was given to 30 teenagers who at least played horror games once. UEQ may provide feedback to developers on the attractiveness, clarity, efficiency, accuracy, stimulation, and novelty aspects of the game. From the results of the UEQ, a reliability test will be carried out using the Cronbach Alpha technique. The results of the descriptive analysis show that these variables are Attractiveness (mean, 0.933), Clarity (mean, 1.808), Efficiency (mean, 1.508), Accuracy (mean, 0.217), Stimulation (mean, 0.667) and Novelty (mean, 0.242). Attractiveness, clarity, and efficiency averaged positive results. The average aspects of accuracy, stimulation, and novelty of the game get neutral results. The results of the reliability test conducted on UEQ data obtained a Cronbach alpha value > 0.6 which indicates that the research data used to test the player experience are considered reliable so that they can be used to provide input for future development of the Alas Tilas game. To increase the average score, the researcher provides recommendations for improvement, namely, adjusting the accuracy and novelty aspects of the horror scenario game entitled Alas Tilas East Java. Therefore, it is expected to improve the quality of the game. Developing a 2D horror game and evaluating the reliability of the player experience are two things that are interrelated and equally important. Developers must ensure that the game can provide a satisfying and reliable gaming experience for its players. This study aims to evaluate the reliability of the player's experience in the game entitled Alas Tilas, East Java. This study used User Experience Questionnaire (UEQ) in Indonesian as survey approach method which was given to 30 teenager respondents who at least played horror games once. UEQ may provide feedback to developers on the Attractiveness, Clarity, Efficiency, Accuracy, Stimulation, and Novelty aspects of the game. From the results of the UEQ, a reliability test will be carried out using the Cronbach Alpha Technique. The results of the descriptive analysis show that these variables are Attractiveness (mean, 0.933), Clarity (mean, 1.808), Efficiency (mean, 1.508), Accuracy (mean, 0.217), Stimulation (mean, 0.667) and Novelty (mean, 0.242). Attractiveness, Clarity and Efficiency averaged positive results. While the average aspects of accuracy, stimulation and novelty of the game get neutral results. The results of the reliability test conducted on UEQ data obtained a Cronbach alpha value > 0.6 which indicates that the research data used in testing player experience is considered reliable so that it can be used to provide input for future Alas Tilas game development. To increase the average score, the researcher provides recommendations for improvement, namely adjusting the Accuracy and Novelty aspects of the horror scenario game entitled Alas Tilas East Java. So that it is expected to improve the quality of the game. 
The Joint Channel Coding and Pre-Distortion Technique on the USRP-Based MIMO-OFDM System
No full textModern wireless communication systems use orthogonal frequency division multiplexing (OFDM), a multi-carrier modulation method that resists multipath channels and provides bandwidth efficiency. OFDM is generally used with a Multiple-Input Multiple-Output (MIMO) system to boost diversity gain and channel capacity. MIMO-OFDM has several advantages, but its high PAPR value is a drawback. A non-linear high-power amplifier (HPA) can distort signals with high PAPR values. This issue can be resolved by employing predistortion, which compensates for nonlinear HPA. In addition to PD, channel coding can be used to improve the quality of systems with high PAPR values by adding redundant bits to the bits to be sent. In this paper, we report the experimental evaluations of the joint channel coding and pre-distortion (PD) technique on a 2x2 MIMO OFDM system using USRP hardware. The experiments are conducted in two scenarios: line-of-sight (LOS) and nonline-of-sight (NLOS) scenarios. The channel coding used in this scenario is convolutional code with code rates of 1/2, 2/3, and 3/4. From the results of the experiment, it can be seen that the system that uses PD combined with the convolution code produces better performance in the LOS and NLOS scenarios compared to the system without PD. In the LOS scenario, the use of PD can improve the SNR value of code rates 1/2, 2/3, and 3/4 by approximately 58.74%, 75.97%, and 96.20%. In the NLOS scenario, the use of PD can improve the SNR value of code rates 1/2, 2/3, and 3/4 by about 60.71%, 73.59%, and 71.84%. The measurement of the LOS scenario gives a better SNR value than the NLOS scenario, with a maximum SNR value of 30.86 dB, while the maximum SNR value of the NLOS scenario is 30.23 dB. This happened because the LOS scenario suffered minimal multipath fading compared to the NLOS scenarioModern wireless communication systems use Orthogonal Frequency Division Multiplexing (OFDM), a multi-carrier modulation method that resists multipath channels and provides bandwidth efficiency. OFDM is generally used with a Multiple-Input Multiple-Output (MIMO) system to boost diversity gain and channel capacity. MIMO-OFDM has several advantages, but its high PAPR value is a drawback. A non-linear high-power amplifier (HPA) can distort signals with high PAPR values. This issue can be resolved by employing pre-distortion (PD), which compensates for non-linear HPA. In addition to PD, channel coding can be used to improve the quality of systems with high PAPR values by adding redundant bits to the bits to be sent. In this paper, we report the experimental evaluations of the joint channel coding and pre-distortion (PD) technique on a 2x2 MIMO OFDM system using USRP hardware. The experiments are conducted in two scenarios: line-of-sight (LOS) and non-line-of-sight (NLOS) scenarios. The channel coding that is used in this scenario is convolutional code with code rates of 1/2, 2/3, and 3/4. From the experiment results, it can be seen that the system using PD combined with the convolution code produces better performance in LOS and NLOS scenarios compared to the system without PD. In the LOS scenario, the use of PD can improve the SNR value of code rates 1/2, 2/3, and 3/4 by about 58.74%, 75.97%, and 96.20%. In the NLOS scenario, the use of PD can improve the SNR value of code rates 1/2, 2/3, and 3/4 by about 60.71%, 73.59%, and 71.84%. The LOS scenario measurement gives a better SNR value than the NLOS scenario, with a maximum SNR value of 30.86 dB, while the NLOS scenario maximum SNR value is 30.23 dB. This happened because the LOS scenario suffered minimal multipath fading compared to the NLOS scenario
The Impact of IoT on The Storing Process of Leather Raw Material
Get PDFA critical step in processing leather raw material is the storing that keeps them in good condition and not easily damaged through the right temperature and humidity level, as otherwise the quality of leather raw material would not be consistent and its economic value would be low. This particular study, therefore, is to conduct an experiment that uses the Internet of Things (IoT) which allows proactive monitoring in keeping a specific temperature and humidity level in their storing process. Hence the subsequent leather processing can be done at the optimal level. The experimentation showed positive results as the use of IoT made storing process of leather raw materials became more proactive and run three times more effective, from just 8 hours to 24 hours, and also brought about a positive effect on the economic efficiency and effectiveness as it enable users to produce more consistent quality of leather raw material whilst the total operating costs remained and even lower. Besides its economic impact, IoT has increased the workers' and their relatives' welfare and social life driven by better income and less time consuming. It also brings about some positive environmental effects as it reduced carbon emissions by keeping energy waste at a lower level.
A critical step in processing leather raw material is the storing that keeps them in good condition and not easily damaged through the right temperature and humidity level, as otherwise the quality of leather raw material would not be consistent and its economic value would be low. This particular study, therefore, is to conduct an experiment that uses the Internet of Things (IoT) which allows proactive monitoring in keeping a specific temperature and humidity level in their storing process. Hence the subsequent leather processing can be done at the optimal level. The experimentation showed positive results as the use of IoT made storing process of leather raw materials became more proactive and run three times more effective, from just 8 hours to 24 hours, and also brought about a positive effect on the economic efficiency and effectiveness as it enable users to produce more consistent quality of leather raw material whilst the total operating costs remained and even lower. Besides its economic impact, IoT has increased the workers' and their relatives' welfare and social life driven by better income and less time consuming. It also brings about some positive environmental effects as it reduced carbon emissions by keeping energy waste at a lower level
Process Mining pada Proses Penerimaan Mahasiswa Baru di Telkom University dengan Genetic Miner
No full textThe selection process for new students at Telkom University, also known as SMB Telkom University has been running for years and already has its process flow. However, the existing process flow can be further improved to better reflect the actual field processes and become more accurate. Process mining can enhance this process flow by creating a new process flow based on event logs or previously executed processes. One of the algorithms in process mining is genetic process mining, where process mining is performed multiple times over several generations and genetic algorithms such as crossover and mutation are applied to generate a more accurate process model compared to other process mining algorithms such as heuristic and inductive mining. After conducting experiments, the best process model that was produced was at the 100th generation which has a fitness point of 0.755910819 and precision point of 0.742857143, after examining the parameters and the resulting Petri net or process flow that was produced it was concluded that the process model obtained from the application of Genetic Process Mining to SMB Telkom University is not very good because the resulting Petri net has several duplicate activities and appears to be non-linear. This could be due to several factors i.e., incompatible, or inaccurate data.Proses seleksi untuk mahasiswa baru di Telkom University telah berjalan selama bertahun-tahun dan sudah memiliki alur proses tersendiri. Namun, alur proses yang ada dapat ditingkatkan lebih lanjut untuk lebih mencerminkan proses lapangan yang sebenarnya dan menjadi lebih akurat. Process mining dapat meningkatkan alur proses ini dengan menciptakan alur proses baru berdasarkan log acara atau proses yang telah dilaksanakan sebelumnya. Salah satu algoritma dalam process mining adalah genetic process mining, di mana process mining dilakukan beberapa kali selama beberapa generasi dan algoritma genetik seperti crossover dan mutasi diterapkan untuk menghasilkan model proses yang lebih akurat dibandingkan dengan algoritma process mining lainnya seperti heuristic dan inductive mining. Dalam penelitian ini, diharapkan bahwa dengan menerapkan genetic process mining pada alur proses seleksi di Telkom University, alur proses yang ada dapat ditingkatkan dan menjadi alur proses yang lebih efektif dan akurat. Setelah melakukan percobaan dan menguji parameter dari model proses yang dihasilkan, seperti kecocokan (fitness) dan kepresisian (precision), kesimpulan yang diambil adalah bahwa model proses yang diperoleh dari aplikasi Genetic Process mining tidak begitu baik karena Petri net yang dihasilkan memiliki beberapa aktivitas yang duplikat dan terlihat tidak linear. Hal ini bisa disebabkan oleh data yang tidak kompatibel atau tidak akurat