Computer Science and Information Technologies (E-Journal)
Not a member yet
149 research outputs found
Sort by
Fine grained irony classification through transfer learning approach
Full text linkNowadays irony appears to be pervasive in all social media discussion forums and chats, offering further obstacles to sentiment analysis efforts. The aim of the present research work is to detect irony and its types in English tweets We employed a new system for irony detection in English tweets, and we propose a distilled bidirectional encoder representations from transformers (DistilBERT) light transformer model based on the bidirectional encoder representations from transformers (BERT) architecture, this is further strengthened by the use and design of bidirectional long-short term memory (Bi-LSTM) network this configuration minimizes data preprocessing tasks proposed model tests on a SemEval2018 task 3, 3,834 samples were provided. Experiment results show the proposed system has achieved a precision of 81% for not irony class and 66% for irony class, recall of 77% for not irony and 72% for irony, and F1 score of 79% for not irony and 69% for irony class since researchers have come up with a binary classification model, in this study we have extended our work for multiclass classification of irony. It is significant and will serve as a foundation for future research on different types of irony in tweets
Exploring application portfolio management in Indonesia: a case study of the Indonesia agency for the assessment and application of technology
Full text linkDue to rapid information and communication technology (ICT) growth, government agencies must manage more digital apps to support public service operations and administration. In Indonesia, at least 400,000 applications have been received by various ministries and government agencies. This amount will hurt the ICT budget’s investment and waste if no approach is employed. Problems can be solved with application portfolio management (APM). Indonesian government agencies’ AMP implementation is unclear. APM is applied at the government research institute agency for the assessment and application of technology (BPPT) in this study. BPPT was chosen due to APM’s lack of ICT investment management. This research examined 41 submissions from the 2019 digital transformation initiative. APM selected two mapping models. The outcome indicates how APM may offer ICT strategies for current applications to government entities. This analysis mapped existing applications into two models: McFarlan’s strategic grid and gartner’s TIME model. Mapping findings from these two models-technical health evaluations and regulatory compliance-may be used for application sustainability suggestions. 11 treatments were advised for maintenance and investment, while 4 applications were recommended for removal. This research helps us understand how the Indonesian government institute maintains its application portfolio and how APM might be a valuable method for application management
Comparative study of ensemble deep learning models to determine the classification of turtle species
Full text linkSea turtles are reptiles listed on the international union for conservation of nature (IUCN) red list of threatened species and the convention on international trade in endangered species of wild fauna and flora (CITES) Appendix I as species threatened with extinction. Sea turtles are nearly extinct due to natural predators and people who are frequently incorrect or even ignorant in determining which turtles should not be caught. The aim of this study was to develop a classification system to help classify sea turtle species. Therefore, the ensemble deep learning of convolutional neural network (CNN) method based on transfer learning is proposed for the classification of turtle species found in coastal communities. In this case, there are five well-known CNN models (VGG-16, ResNet-50, ResNet-152, Inception-V3, and DenseNet201). Among the five different models, the three most successful were selected for the ensemble method. The final result is obtained by combining the predictions of the CNN model with the ensemble method during the test. The evaluation result shows that the VGG16 - DenseNet201 ensemble is the best ensemble model, with accuracy, precision, recall, and F1-Score values of 0.74, 0.75, 0.74, and 0.76, respectively. This result also shows that this ensemble model outperforms the original model
Hybrid transformation with 1’st order statistics for medical image analysis and classification
Full text linkSkin cancer, one of the most critical forms of cancer, required early detection and documentation for efficient treatment, especially as certain types are fatal. In this study, an artificial neural network (ANN) was utilized to discover and index diverse melanomas using the ISIC 2018 dataset. The pre-processing phase is stringent as it insulates the cancerous fraction of a skin image, involving removing, trimming, thinning, and normalizing. In this phase, unwanted hair pieces on the image are eliminated in this phase. Feature extraction from the clipped image is achieved using a discrete cosine transform (DCT) and a gradient transform to transform it into frequency domain coefficients. Statistical feature extraction is used to reduce the amount of data required for ANN training. A dataset from ISIC 2018 that consists of seven different images from dermoscopic procedures for classification purposes is used in the empirical investigation. An accuracy of 85.44% for DCT in the sub-bands and 76.07% for the sub-band gradient transform was achieved by the applied ANN. The hybrid system's mean squared error (MSE) was discovered to be 3.52×10-4. The work highlights the potential of ANN in the early detection of skin cancer, supporting more efficient treatment and preventing advanced cases
Investigating the impact of data scaling on the k-nearest neighbor algorithm
Full text linkThis study investigates the impact of data scaling techniques on the performance of the k-nearest neighbor (KNN) algorithm using ten different datasets from various domains. Three commonly used data scaling techniques, min-max normalization, Z-score, and decimal scaling, are evaluated based on the KNN algorithm's performance in terms of accuracy, precision, recall, F1-score, runtime, and memory usage. The study aims to provide insights into the applicability and effectiveness of different scaling techniques in different contexts, aid in the design and implementation of machine learning systems, and help identify the strengths and weaknesses of each technique and their suitability for specific types of data. The results show that data scaling significantly affects the performance of the KNN algorithm, and the choice of scaling method can have significant implications for practical applications. Moreover, the performance of the three scaling techniques varies across different datasets, suggesting that the choice of scaling technique should be made based on the specific characteristics of the data. Overall, this study provides a comprehensive analysis of the impact of data scaling techniques on the KNN algorithm's performance and can help practitioners and researchers in the machine learning community make informed decisions when designing and implementing machine learning systems
Net impact implementation application development life-cycle management in banking sector
Full text linkDigital transformation in the banking sector creates a lot of demand for application development, either new development or application enhancement. Continuous demand for reimagining, revamping, and running applications reliably needs to be supported by collaboration tools. Several big banks in Indonesia use Atlassian products, including Jira, Confluence, Bamboo, Bitbucket, and Crowd, to support strategic company projects. We need to measure the net impact of application development life-cycle management (ADLM) as a collaboration tool. Using the deLone and McLean model, process questionnaire data from banks in Indonesia that use ADLM. Processing data using structural equation modeling (SEM), multiple variables are analyzed statistically to establish, estimate, and test the causation model. The conclusions highlight that system quality strongly affected only User Satisfaction (p=0.049 and β=0.39). Information quality strongly affected use (p=0.001 and β=0.84) and strongly affected user satisfaction (p=0.169 and β=0.28). Service quality strongly affected only use (p=0.127 and β=0.31). Conclusion research verifies the information system's achievement approach described by DeLone and McLean. Importantly, it was discovered that system usability and quality were key indicators of ADLM success. To fulfill their objective, ADLM must be developed in a way that is simple to use, adaptable, and functional
Caring jacket: health monitoring jacket integrated with the internet of things for COVID-19 patients
Full text linkOne of the policies that have been made by the World Health Organization (WHO) and the Indonesian government during this COVID-19 pandemic, is to use an oximeter for self-isolation patients. The oximeter is used to monitor the patient if happy hypoxia which is a silent killer, happens to the patient. To maintain body endurance, exercise is needed by COVID-19 patients, but doing too much exercise can also cause decreased immunity. That’s why fatigue level and exercise intensity need to be monitored. When exercising, social distancing protocol should be also reminded because can lower COVID-19 spreading up to 13.6%. To solve this issue, the Caring Jacket is proposed which is a health monitoring jacket integrated with an IoT system. This jacket is equipped with some sensors and the global positioning system (GPS) for tracking. The data from the test showed the temperature reading accuracy is up to 99.38%, the oxygen rate up to 97.31%, the beats per minute (BPM) sensor up to 97.82%, and the precision of all sensors is 97.00% compared with a calibrated device
The antecedent e-government quality for public behaviour intention, and extended expectation-confirmation theory
Full text linkThe main objective of the study is to identify the antecedent of leadership quality, public satisfaction and public behaviour intention of e-government service. Also, this study integrated e-government quality to expectation-confirmation model. In order to achieve these goals, observational research was then carried out to collect primary information, using the method of data dissemination and obtaining the opinion of 360 from the public using the e-government service and some of the e-government and software quality experts. The results of the study show that the positive association among the e-government services quality and public perceived usefulness, public expectation confirmation, leadership quality and public satisfaction that also play a positive role on the public behavior intention
Designing a framework for blockchain-based e-voting system for Libya
Full text linkA transition to democratic rule is considered the first step down a long road towards Libya’s recovery and prosperity. Thus, it strives to improve the country’s elections by introducing new technologies. A blockchain is a distributed ledger that is characterised by independence and security. Therefore, it has been widely applied in various fields ranging from credit encryption and digital currency. With the development of internet technology, electronic voting (E-voting) systems have been greatly popularised. However, they suffer from various security threats, which create a sense of distrust among existing systems. Integrating blockchain with online elections is a promising trend, which could lead to make an election transparent, immutable, reliable, and more secure. In this paper, we present a literature review and a case analysis of blockchain technology. Moreover, a framework for an E-voting system based on blockchain is proposed. The methodology is adopted on the basis of three activities, they are identification of the relevant literature about E-voting, system modelling, and the determination of suitable technological tools. The framework is secure and reliable. Thus, it could help increase the number of voters and ensure a high level of participation, as well as facilitate free and fair electoral processes
The observed preprocessing strategies for doing automatic text summarizing
Full text linkIt is challenging for humans to keep up with the rapid creation of digital information due to the explosion of digital information. A written document can be analyzed to extract meaningful information using automatic text summarization. This research proposes 16 different experimental settings in which the model developed by IndoBERT will be applied in order to answer the question of how much of an impact preprocessing has on the quality of summaries produced by automatic text summarization. In order to answer this question, the researchers have devised this study. In this study, we will explicitly talk about preprocessing strategies by conducting tests with different combinations of preprocessing techniques. These techniques include data cleansing, stopwords, stemming, and case folding. After that, the recall-oriented understudy for gisting evaluation (ROUGE) assessment will be used to conduct the measurement of the research results. According to the findings of this research, the optimal level of performance may be accomplished by combining the processes of data cleaning and case folding with scores of 0.78, 0.60, and 0.68 for ROUGE-1, ROUGE-2, and ROUGE-L respectivel