Kinetik: Game Technology, Information System, Computer Network, Computing, Electronics, and Control
Not a member yet
424 research outputs found
Sort by
Enhancing Qur'anic Recitation Experience with CNN and MFCC Features for Emotion Identification
Get PDFIn this study, MFCC feature extraction and CNN algorithms are used to examine the identification of emotions in the murottal sounds of the Qur'an. A CNN model with labelled emotions is trained and tested, as well as data collection of Qur'anic murottal voices from a variety of readers using MFCC feature extraction to capture acoustic properties. The outcomes show that MFCC and CNN work together to significantly improve emotion identification. The CNN model attains an accuracy rate of 56 percent with the Adam optimizer (batch size 8) and a minimum of 45 percent with the RMSprop optimizer (batch size 16). Notably, accuracy is improved by using fewer emotional parameters, and the Adam optimizer is stable across a range of batch sizes. With its insightful analysis of emotional expression and user-specific recommendations, this work advances the field of emotion identification technology in the context of multitonal music
The Application of PROMETHEE Method in Determining Scholarship Recipients at University
Get PDFThis study aims to use PROMETHEE method as a decision support system in determining the recipients of the Academic Achievement Improvement Scholarship at Universitas Dharmas Indonesia (UNDHARI). The methodological steps include problem identification, analysis, goal setting, and the application of PROMETHEE method. In this study, the criteria and alternatives have been identified to evaluate the scholarship recipients. The criteria weights are set, and the criteria preference types are determined. After obtaining the baseline data from the questionnaire assessments, pairwise preference values and multicriteria preference index values are calculated. Then, the rankings are compiled using Leaving Flow, Entering Flow, and Net Flow methods, resulting in the priority order of the scholarship recipients. The ranking results show that alternative 3 (IS) has the highest Net Flow value (0.30), while alternative 2 (AV) has the lowest Net Flow value (-0.35). Thus, the priority order from highest to lowest is IS, AV, RD, YM, and AV. In the context of Net Flow scores, these results indicate that alternative 3 (IS) has the greatest chance of receiving the academic achievement improvement scholarship. This study provides important insights for UNDHARI in the scholarship recipient determination process using the PROMETHEE method as a decision-making tool
Content-based Filtering Movie Recommender System Using Semantic Approach with Recurrent Neural Network Classification and SGD
Get PDFThe application of recommendation systems has been applied in various types of platforms, especially applications for watching movies such as Netflix and Disney+. The recommendation system is purposed to make it easier for users, especially in choosing a movie because currently the number of movie productions is increasing every day. This research proposed a CBF movie recommendation system by comparing the performance of several semantic methods to be able to get the best rating prediction results. In order to improve the performance quality to get the best rating prediction results, this research utilized semantic feature methods by comparing the performance of the evaluation results produced by the TF-IDF method and word embedding applications, such as BERT, GPT-2, RoBERTa, and implemented RNN model to classify the results of rating prediction. The data were used to generate the recommendation system by involving 854 data movie and 39 accounts with a total of 34,056 movie reviews on Twitter. This research has succeeded in getting a method that produced rating predictions, namely RoBERTa. In the classification process with the RNN model and SGD optimization, the measurement results with confusion matrix by classifying the RoBERTA rating prediction obtained an evaluation value of 0.6514 loss, 95.59% accuracy, and 0.6514 precision
Movie Recommender System on Twitter Using Weighted Hybrid Filtering and GRU
Get PDFThe development of the industry in the film sector has experienced rapid growth, marked by the emergence of film streaming platforms such as Netflix and Disney+. With the abundance of available films, users face difficulty in choosing films that suit their preferences. Recommender systems can be a solution to this problem for users. Recommender systems rely on user reviews, making Twitter a platform that can be used to collect user reviews of a film. This study will develop a recommender system that has the potential to provide item recommendations to users using the weighted hybrid filtering and GRU methods. The weighted hybrid filtering used is a combination of collaborative filtering and content-based filtering methods. The dataset used in this study was obtained by crawling tweets relevant to the feedback of specific accounts regarding a film. The dataset resulting from the data crawling consists of a total of 854 films, 45 users and 34,086 tweets consisting of film reviews from Twitter users. The GRU model classification is performed on the results of weighted hybrid filtering with model optimization involving testing various test size scenarios and optimizer methods. The test sizes used are 40%, 30%, and 20%. The optimizer methods used include Adam, Nadam, Adamax, Adadelta, Adagrad, and SGD. The research results show that the optimal outcome is obtained using the Nadam optimization method. The performance evaluation yielded 85.74% precision, 88.63% recall, 88.63% accuracy, and 86.30% F1-score
Design of SEPIC Converter for Battery Charging System using ANFIS
Get PDFRechargeable batteries are the most widely used medium for storing energy today. One type of rechargeable battery that is widely used is lithium-ion batteries. The large use of lithium-ion batteries in society requires companies to conduct research so that the life time of these batteries can last a long time and charging can take place quickly. Charging system at this time is less efficient in charging lithium batteries where the time needed is still quite long where when lithium batteries are charged with a long time can cause the battery to heat up quickly and can reduce the life time of the battery. To overcome this, a system is needed that can control the battery charger process so that the output voltage and current are constant and battery charging is faster. It is hoped that the SEPIC converter system can help many people who forget to unplug the power supply during the charging process so as to maintain the life time of the battery. Setting the output voltage and current in the DC-DC converter can be done using an Adaptive Neuro Fuzzy Inference System which aims to keep the output of SEPIC stable according to the setting point. In this system, the DC-DC converter used is a SEPIC converter which can increase and decrease the output voltage for battery charging. The battery charging process uses the CC-CV method. In the test, the average error is 0.025% where when the SOC is 60% to 80% the average error is 0.04% and when the SOC is 80% to 95% the average error is 0.0005%
Exploring Trust, Privacy, and Security in Cloud Storage Adoption among Generation Z: An Extended TAM Approach
Get PDFThe incorporation of cloud storage technology holds the promise of significantly enhancing efficiency in various sectors, particularly from the perspective of Generation Z, a demographic known for its meticulous consideration of technology acceptance factors, especially security. This research thoroughly examines the level of acceptance of cloud storage technology among Generation Z. By augmenting the Technology Acceptance Model (TAM) with five core factors and introducing three novel factors—Perceived Security, Perceived Privacy, and Trust—this study not only adheres to traditional acceptance models but also ventures into uncharted territories, marking a significant contribution to understanding technology acceptance. This study meticulously collected data from 408 Generation Z respondents who actively use cloud storage technology, employing an innovative questionnaire disseminated via an online platform. Through sophisticated PLS-SEM data analysis, the study confirmed the positive and significant impact of all tested hypotheses, underscoring the importance of attitudes, perceived benefits, and usability in fostering the intention to use cloud storage. Notably, the added dimensions of privacy and security emerged as critical in enhancing users' trust in cloud storage solutions. Furthermore, this study paves the way for future explorations into technology acceptance across diverse populations and settings, underscoring the critical role of security and privacy in shaping technology adoption decisions among emerging generations
Entropy-Based Feature Extraction and K-Nearest Neighbors for Bearing Fault Detection
Get PDFBearing failures in rotating machines can lead to significant operational challenges, causing up to 45-55% of engine failures and severely impacting performance and productivity. Timely detection of bearing anomalies is crucial to prevent machine failures and associated downtime. Therefore, an approach for early bearing failure detection using entropy-based machine learning is proposed and evaluated while combined with a classifier based on K-Nearest Neighbors (KNN) and Support Vector Machine (SVM). Entropy-based feature extraction should be able to effectively capture the intricate patterns and variations present in the vibration signals, providing a comprehensive representation of the underlying dynamics. The results of the classification carried out by KNN-Entropy have an accuracy value of 98%, while the SVM-Entropy model has an accuracy of 96%. Hence, the Entropy-based feature extraction giving the best accuracy when it is coupled with KNN
The Citizens' Satisfaction on Service Quality of Mobile Government (Case Study : Wargaku Surabaya Application)
Get PDFThis research aims to analyze the citizens’ satisfaction with the service quality of a mobile Government application named WargaKu Surabaya, by using the Mobile Government Service Quality (SQ-mGov) measurement involving four indicators, namely connectivity, interactivity, authenticity, and understandability. The research method used was quantitative using SmartPLS version 0.3 software with primary data sources in the form of questionnaires totaling 100 respondents. This research is interesting because it discusses citizens’ satisfaction with the quality of government services using parameters for measuring the quality of mobile government services. The research results show that the connectivity and understandability influence the citizens’ satisfaction with the service quality of WargaKu Surabaya application. However, the interactivity and authenticity do not affect the citizens’ satisfaction on WargaKu Surabaya application. The practical implications of this research can be used as an input for the government to improve the quality of mobile government services, particularly WargaKu Surabaya application, with the hope that as the service quality mobile Government increases, the citizens’ satisfaction will also improve
Comparative Study of Classification of Eye Disease Types Using DenseNet and EfficientNetB3
Get PDFThe purpose of this research is to build a classification model that can perform the eye disease identification process so that the diagnosis of eye disease can be known and medical action can be taken as early as possible. This research used a dataset which has a total of 4217 eye image data and had 4 main classes namely cataract, diabetic retinopathy, glaucoma, and normal. With the data distribution of 1038 cataract images, 1098 diabetic retinopathy images, 1007 glaucoma images, and 1074 normal images, which of this data will be divided with a data percentage scheme of 50:10:40, 60:10:30, and 70:10:20, to see the results of which dataset division can produce optimal accuracy. In this study, the classification process will use 2 CNN transfer learning architectures, namely DenseNet, and efficientnetb3, which are both trained using the ImagiNet dataset. The results obtained after completing the testing process on the model built using the DenseNet architecture get optimal accuracy when using data division as much as 60:10:30, which is 78.59% while using the efficientnetb3 architecture optimal accuracy results when using the data division of 70:10:20, which is 95.66%. In research on the classification that had previously been done, it is very rare to find a classification process for eye disease types, therefore, in this study, the classification process will be carried out and provide an overview of the eye disease classification process with the CNN transfer learning method with more optimal accuracy results
Pattern Recognition of Bima Script Handwritting using Convolutional Neural Network Method
Get PDFBima is one of the regions in West Nusa Tenggara Province. The Bima script is a cultural heritage used as a means of communication by the Bima community in the past. The decline in the use of the Bima script threatens cultural heritage. The government has addressed this issue by providing training to teachers to teach it in schools, but this has still been insufficient due to the limited number of teachers participating in the training. Therefore, one efficient method to assist with this issue is by leveraging modern technology, particularly through machine learning for handwriting recognition. This study aims to find the best CNN model for recognizing the Bima script with diacritics to help preserve Bima's cultural heritage through handwriting recognition. The CNN model is combined with hyperparameter tuning, and then testing is conducted in four different scenarios to evaluate the performance of each model architecture and hyperparameter variation to find the best combination. The dataset used is sourced from the Kaggle platform, and augmentation is performed to increase the total number of images to 6,750, with each image containing 75 images in 90 different classes. In this study, testing is done by dividing the dataset into training and testing sets in an 80:20 ratio. The test results show high performance, achieving an accuracy of 98.00%, precision of 98.19%, recall of 98.00%, and f1-score of 98.00% in scenario 4