International Journal of Informatics and Communication Technology (IJ-ICT)
Not a member yet
494 research outputs found
Sort by
A routine immunization decision support system framework for vaccine demand forecasting in the city health office
Get PDF“Immunization” has been documented as one of the most flourishing measures for community well-being ever devised. Management of “immunization” information will ensure that children and newborns receive immunization on schedule. However, managing this immunization information is done manually. Customary data processing method are timeintensive, lengthy, slow in progress and susceptible to inaccuracies during encoding, verification, and re-ordering. In this study, a web-based routine immunization decision support system (RIDSS) was conceptualized to address these challenges. The web-based system is an innovative tool designed to streamline vaccine demand forecasting within the city health office (CHO) of Panabo. This system uses time series analysis and machine learning models to output accurate predictions of future vaccination demand. Using historical data on the performance of routine immunization (RI), it allows identification and analysis of actionable signals to facilitate betterinformed decisions with respect to vaccine procurement, distribution and allocation. The system is a substantial improvement of the current basic vaccine supply management, making it possible for Panabo CHO to have an organized program in administering immunization. Key stakeholders identified were presented with the prototype of system to assure effectiveness and utility. An act of major recognition to the system and its relevance in community health
Creating a smart bedroom for children by connecting PIR and LDR sensors to a microcontroller Arduino UNO ATmega328P
Get PDFIntelligent electronic systems are increasingly prevalent in modern society. The development of smart bedrooms for young children, especially those with developmental disabilities, it is based on the responses of passive infrared (PIR) and light dependent resistor (LDR) sensors. The PIR sensor detects children’s movement during the night, triggering the microcontroller to send a bit of 1 to the microcontroller pin connected to an electromagnetic relay, which then switches on a 220 VAC light to illuminate the bedroom. This only occurs if the LDR sensor has high resistance, indicating that the environment is completely dark. The functionality of this intelligent system mainly depends on the program code (sketch) uploaded to the Arduino UNO microcontroller module. The microcontroller is programmed to perform specific functions based on the sensors data. It is based on the responses of PIR and LDR sensors. The PIR sensor detects children’s movement during the night, triggering the microcontroller to send a bit of 1 to the microcontroller pin connected to an electromagnetic relay, which then switches on a 220 VAC light to illuminate the bedroom. This only occurs if the LDR sensor has high resistance, indicating that the environment is completely dark
A hybrid framework for enhanced intrusion detection in cloud environments leveraging autoencoder
Get PDFIn today’s world, the significance of network security and cloud environments has grown. The rising demand for data transmission, along with the versatility of cloud-based solutions and widespread availability of global resources, are key drivers of this growth. In response to rapidly evolving threats and malicious attacks, developing a robust intrusion detection system (IDS) is essential. This study addresses the imbalanced data and utilizes an unsupervised learning approach to protect network data. The suggested hybrid framework employs the CIC-IDS2017 dataset, integrating methods for handling imbalanced data with unsupervised learning to enhance security. Following preprocessing, principal component analysis (PCA) reduces the dimensionality from eighty features to twenty-three features. The extracted features are input into density-based spatial clustering of applications with noise (DBSCAN), a clustering algorithm. particle swarm optimization (PSO) optimizes DBSCAN, grouping similar traffic and enhancing classification. To address the imbalances in the learning process, the autoencoder (AE) algorithm demonstrates unsupervised learning. The data from the cluster is input into the AE, a deep learning algorithm, which classifies traffic as normal or an attack. The proposed approach (PCA+DBSCAN+AE) attains remarkable intrusion detection accuracy exceeding 98%, and outperforms five contemporary methodologies
Indonesian automated short-answer grading using transformers-based semantic similarity
Get PDFAutomatic short answer grading (ASAG) systems offer a promising solution for improving the efficiency of reading literacy assessments. While promising, current Indonesian artificial intelligence (AI) grading systems still have room for improvement, especially when dealing with different domains. This study explores the effectiveness of large language models, specifically bidirectional encoder representations from transformers (BERT) variants, in conjunction with traditional hand-engineered features, to improve ASAG accuracy. We conducted experiments using various BERT models, hand-engineered features, text pre-processing techniques, and dimensionality reduction. Our findings show that BERT models consistently outperform traditional methods like term frequency-inverse document frequency (TF-IDF). IndoBERTLite-Base-P2 achieved the highest quadratic weighted kappa (QWK) score among the BERT variants. Integrating handengineered features with BERT resulted in a substantial enhancement of the QWK score. Utilizing comprehensive text pre-processing is a critical factor in achieving optimal performance. In addition, dimensionality reduction should be carefully used because it potentially removes semantic information
Legal challenges of artificial intelligence in the European Union’s digital economy
Get PDFThis article critically examines the legal and regulatory challenges posed by artificial intelligence (AI) within the European Union’s (EU) digital economy, focusing on the recently adopted EU Ai Act (Regulation 2024/1689). While previous studies have addressed AI's ethical and theoretical dimensions, this research fills a gap by analyzing the Act’s practical application across its temporal, personal, material, and territorial scopes. The study adopts a qualitative legal methodology, integrating doctrinal interpretation, comparative analysis, and systemic review of EU and international legal instruments. Key findings reveal that the EU AI Act establishes a pioneering risk-based regulatory framework, distinguishing itself through strong safeguards for fundamental rights, transparency, and human oversight. However, critical limitations remain, including ambiguous high-risk classifications, reliance on provider self-assessment, and exemptions for national security that may undermine human rights protections. The article compares the EU approach with those of the United States and China, illustrating divergent models of AI regulation and their global implications. It argues that while the EU AI Act sets an ambitious precedent, its success depends on effective enforcement, stakeholder compliance, and international regulatory convergence. By addressing these challenges, the EU can shape a globally influential framework for ethical and responsible AI deployment. This study contributes to the evolving academic and policy debate on AI governance by offering practical insights and recommendations for future research and legal development
Pneumonia detection system using convolutional neural network with DenseNet201 architecture
Get PDFThe diagnosis of pneumonia remains a significant challenge for medical practitioners worldwide, particularly in regions with limited healthcare resources. Traditional interpretation of chest X-rays is time-consuming and often subjective, especially when images are of low quality. This study presents the development of a web-based system utilizing the DenseNet201 architecture to address these challenges. A series of experiments were conducted to evaluate three optimizers Adam, Adamax, and Adadelta over fifty epochs. Among them, Adamax yielded the best performance, achieving a training accuracy of 93.67% and a validation accuracy of 94.20%. When tested on new data, the system consistently delivered high performance, with accuracy, precision, recall, and F1 score all reaching 96%. These results suggest that the proposed system has the potential to significantly enhance the accuracy and efficiency of pneumonia diagnosis based on chest X-rays
Efficient design of approximate carry-based sum calculating full adders for error-tolerant applications
Get PDFApproximate computing is an innovative circuit design approach which can be applied in error-tolerant applications. This strategy introduces errors in computation to reduce an area and delay. The major power-consuming elements of full adder are XOR, AND, and OR operations. The sum computation in a conventional full adder is modified to produce an approximate sum which is calculated based on carry term. The major advantage of a proposed adder is the approximation error does not propagate to the next stages due to the error only in the sum term. The proposed adder was coded in verilog HDL and verified for different bit sizes. Results show that the proposed adder reduces hardware complexity with delay requirements
Review of NLP in EMR: abbreviation, diagnosis, and ICD classification
Get PDFThis review explores state-of-the-art natural language processing (NLP) methods applied to electronic medical records (EMRs) for key tasks such as expanding medical abbreviations, automated diagnosis generation, international classification of diseases (ICD) classification, and explaining model outcomes. With the growing digitization of healthcare data, the complexity of EMR analysis presents a significant challenge for accurate and interpretable results. This paper evaluates various methodologies, highlighting their strengths, limitations, and potential for improving clinical decision-making. Special attention is given to abbreviation expansion as a crucial step for disambiguating terms in the clinical text, followed by an exploration of auto-diagnosis models and ICD code assignment techniques. Finally, interpretability methods like integrated gradients and attention-based approaches are reviewed to understand model predictions and their applicability in healthcare. This review aims to provide a comprehensive guide for researchers and practitioners interested in leveraging NLP for clinical text analysis
An artificial intelligent system for cotton leaf disease detection
Get PDFThis study aims to develop a deep learning-based system for the detection and classification of diseases in cotton leaves, with the goal of aiding in early diagnosis and disease management, thereby enhancing agricultural productivity in India. The study utilizes a dataset of cotton leaf images, classified into four categories: Fusarium wilt, Curl virus, Bacterial blight, and Healthy leaves. The dataset is used to train and evaluate various CNN models such as basic CNN, VGG19, Xception, InceptionV3, and ResNet50. These models were evaluated on their accuracy in identifying the presence of diseases and classifying cotton leaf images into the respective categories. The models were trained using standard deep learning frameworks and optimized for high performance. The results indicated that ResNet50 achieved the highest accuracy of 100%, followed by InceptionV3 with 98.75%, and VGG19 and Xception both with 97.50%. The basic CNN model showed an accuracy of 96.25%. These models demonstrated strong potential for accurate multi-class classification of cotton leaf diseases. This study emphasizes the potential of deep learning in agricultural diagnostics. Future research can focus on improving model robustness, incorporating larger datasets, and deploying the system for real-time field use to assist farmers in disease management and improving cotton production
A custom-built deep learning approach for text extraction from identity card images
Get PDFInformation found on an identity card is needed for different essential tasks and manually extracting this information is time consuming, resource exhaustive and may be prone to human error. In this study, an optical character recognition (OCR) approach using deep learning techniques is proposed to automatically extract text related information from the image of an identity card in view of developing an automated client onboarding system. The OCR problem is divided into two main parts. Firstly, a custom-built image segmentation model, based on the U-net architecture, is used to detect the location of the text to be extracted. Secondly, using the location of the identified text fields, a (CRNN) based on long short-term memory (LSTM) cells is trained to recognise the characters and build words. Experimental results, based on the national identity card of the Republic of Mauritius, demonstrate that our approach achieves higher accuracy compared to other studies. Our text detection module has an intersection over union (IOU) measure of 0.70 with a pixel accuracy of 98% for text detection and the text recognition module achieved a mean word recognition accuracy of around 97% on main fields of the identity card