JOIV : International Journal on Informatics Visualization
Not a member yet
786 research outputs found
Sort by
A Review on Classifying and Prioritizing User Review-Based Software Requirements
Get PDFUser reviews are a valuable source of feedback for software developers, as they contain user requirements, opinions, and expectations regarding app usage, including dislikes, feature requests, and reporting bugs. However, extracting and analyzing user requirements from user reviews is ineffective due to the large volume, unstructured nature, and varying quality of the reviews. Therefore, further research is not just necessary but crucial to effectively explore methods to gather informative and meaningful user feedback. This study aims to investigate, analyze, and summarize the methods of requirement classification and prioritization techniques derived from user reviews. This review revealed that leveraging opinion mining, sentiment analysis, natural language processing, or any stacking technique can significantly enhance the extraction and classification processes. Additionally, an updated matrix taxonomy has been developed based on a combination of definitions from various studies to classify user reviews into four main categories: information seeking, feature request, problem discovery, and information giving. Furthermore, we identified Naive Bayes, SVM, and Neural Networks algorithms as dependable and suitable for requirement classification and prioritization tasks. The study also introduced a new 4-tuple pattern for efficient requirement prioritization, which included elicitation technique, requirement classification, additional factors, and higher range priority value. This study highlights the need for better tools to handle complex user reviews. Investigating the potential of emerging machine learning models and algorithms to improve classification and prioritization accuracy is crucial. Additionally, further research should explore automated classification to enhance efficiency
A Comprehensive Review on Cancer Detection and Classification using Medical Images by Machine Learning and Deep Learning Models
Get PDFIn day-to-day life, machine learning and deep learning plays a vital role in healthcare applications to predict various diseases such as cancer, heart attack, mental problem, Parkinson, etc. Among these diseases, cancer is the life-threatening disease that leads a human being to death. The primary aim of this study is to provide a quick overview of various cancers and provides a comprehensive overview of machine learning and deep learning techniques in the detection and classification of several types of cancers. The significance of machine learning and deep learning in detecting various cancers using medical images were concentrated in this study. It also discusses various machine learning and deep learning algorithms that lead to accurate classification of medical images, early diagnosis, and immediate treatment for the patients and explores the methodologies which has been used to predict the cancer with the help of low dose computer tomography to reduce cancer related deaths. As the study narrows down the research into lung cancer, it combats the findings limitations in lung cancer detection models and highlights the need for a deep study of novel cancer detection algorithms. In addition, the review also finds the role of setting up data in lung cancer and the potential of genetic markers in stabilizing the accuracy of machine learning models. Overall, this study gives valuable suggestions to achieve more accuracy in cancer detection and classification using machine learning and deep learning techniques.
Involvement of Various Selection Methods for Genetic Algorithms in Determining the Optimal Production Schedule Problem
Get PDFThis research investigates using genetic algorithms (GA) to optimize production scheduling in Medan's shoe industry. The study compares traditional manual and First Come First Serve (FCFS) methods against a GA approach, incorporating selection variations such as Boltzmann, Fitness Uniform Selection Scheme (FUSS), Exponential Rank Selection, and Roulette Wheel Selection. The optimal production order is derived from the chromosome with the highest fitness. Results indicate that GA with FUSS selection significantly reduces production time from 73,630 minutes to 45,650 minutes, achieving a 35% improvement in efficiency. This optimization is attributed to FUSS’s ability to maintain a diverse population, preventing premature convergence and ensuring a broader solution for space exploration. Additionally, it was found that using a smaller population size relative to the number of generations yields better optimization results. The study also demonstrates that while Roulette Wheel Selection shows more variability, it achieves higher optimization over time than FCFS. The practical implications of these findings are substantial for the shoe industry, including faster production cycles, better resource allocation, and an enhanced ability to meet customer demands. These benefits are exemplified by implementing the SISPROMA application, an innovative production scheduling information system that leverages machine learning to optimize scheduling in the manufacturing industry. This study provides valuable insights into applying genetic algorithms for production scheduling, highlighting their potential to enhance operational efficiency and reduce costs. Future research should explore additional optimization techniques and real-world applications to validate and extend these findings, ensuring broader applicability and continuous improvements in manufacturing efficiency
Diagnosis of Diseases in Rubber Stems Using the Dempster Shafer Method
Get PDFRubber (Hevea Brasiliensis) is a non-timber forest product originating from the Americas and is currently widely distributed worldwide, including in East Kalimantan, Indonesia. In their management in East Kalimantan, farmers often encounter diseases in rubber plants, especially diseases of the stems, which can cause plant death. This disease requires treatment, but if it is too severe, it can harm farmers economically and in production, so it is essential for farmers to recognize the symptoms of this disease early from changes in the rubber plant stems. This study aims to diagnose diseases of rubber stems using the Dempster Shafer method. Dempster Shafer is a relevant method for overcoming the uncertainty of symptoms and rules, enabling expert systems to generate conclusions with certainty. This method has advantages in solving various problems and simultaneously combining evidence (facts) from several sources. This research was conducted by analyzing a dataset of 80 data, covering 7 types of diseases and 27 different symptoms. The accuracy test results show that the research has an accuracy rate of 96.25%. The implications of this research are significant. It is hoped that it can significantly help rubber plantation farmers in East Kalimantan and also make a valuable contribution to agricultural and plantation extension agents in overcoming the challenges faced due to diseases in rubber plant stems. Thus, this research could increase the productivity and sustainability of the rubber plantation sector in this region
Collaborative Intrusion Detection System with Snort Machine Learning Plugin
Get PDFThe increasing prevalence of cybercrime and cyber-attacks underscores the imperative need for organizations to implement robust network security measures. Nevertheless, current Intrusion Detection Systems (IDS) often rely on single-sensor or multi-sensor in the same type of IDS, including Host-Based IDS (HIDS) or Network-Based IDS (NIDS), which inherently possess limited detection capabilities. To address this limitation, this research combines NIDS and HIDS components into a collaborative-IDS system, thus expanding the scope of intrusion detection and enhancing the efficacy of the established attack mitigation system. However, the integration of NIDS and HIDS introduces formidable challenges, notably the elevated rates of False Positive and False Negative alerts. To surmount these challenges, the researcher employs machine learning techniques in the form of Snort plugins and comparison methods to heighten the precision of attack detection. The obtained results unequivocally illustrate the effectiveness of this approach. Using a Support Vector Machine for static analysis of the NSL-KDD dataset attains an outstanding 99% detection rate for Denial of Service (DoS) attacks and an impressive 98% detection rate for Probe attacks. Furthermore, in dynamic real-time attack simulations, the machine learning plugins exhibit remarkable proficiency in detecting various types of DoS attacks, concurrently offering more comprehensive identification of SYN Flooding DoS attacks compared to the Snort community rules set. These findings signify a significant advancement in intrusion detection, paving the way for more robust and accurate network security systems in an era of escalating cyber threats
Recipient Feasibility Decision Support System Micro Small Medium Business Assistance Use Method Analytic Hierarchy Process and Simple Additives Weighting
Get PDFStudy This aims To determine the eligibility of MSME assistance recipients with the method AHP (Analytic Hierarchy Process) And SAW (Simple Additives weighting). The AHP method is used to determine the weight of each criterion. Meanwhile, SAW is used To determine the rank selection of beneficiaries. This is very important for Indonesia's economy during the crisis, Where MSME's own Power stands to face a crisis economy. Criteria used in a way This uses six measures: type of business, Amount of power Work, turnover per month, amount of assets, sector MSME, And sector business. Decision support systems are designed to support someone who must make certain decisions. That is, interactive, Flexible, Data quality, and Expert Procedure. Study System Supporters Decision Appropriateness Recipient Help Business Micro Community Use Analytic Hierarchy Process (AHP) and Simple Additive Methods weighting (SAW), Study This done in Subdistrict Intersection Three Regency Pidie Aceh Province to facilitate the Selection of Eligibility of Government Assistance Recipients For Build a business Micro Society. Testing is done in this study, namely black box testing. Results Testing black box shows that the system can walk with Good by function, with results calculation method AHP and results calculation method SAW in determining eligibility selection MSME aid recipients. The results of the level of accuracy testing on the AHP and SAW methods with six criteria and alternatives the requirements is 75%
Optimizing Linked List-based Smart Contract on Ethereum with IPFS for E-book Management System
Get PDFPeople are now widely adopting digital assets in various applications, integrating them into almost every aspect of their lives. Electronic books, or e-books, are one of the digital assets that result from the transformation of physical reading material into the digital world. Nowadays, blockchain is used in many industries because it provides immutable and transparent records. E-book publishers may take this opportunity to adopt blockchain technology for e-book data management. However, blockchain storage is limited; thus, storing the e-book files in blockchain is not recommended. A decentralized storage system, such as InterPlanetary Files Systems (IPFS), is an alternative way to store large files like e-books. IPFS can facilitate the storage of e-book files while the metadata is stored in the blockchain. The e-book metadata should be stored in a structured way for effective search and retrieval. E-book metadata could be added, deleted, and updated occasionally. Nevertheless, some data structures often struggle with dynamic collections of records. This paper proposes a linked list-based smart contract on Ethereum that integrates with IPFS for the e-book management system. We demonstrate the implementation of a linked list smart contract for insertion, deletion, update, retrieval, and traversal of the e-book’s metadata. The result shows that a linked list-based smart contract with IPFS could offer a robust solution for e-book data management. This solution provides more opportunities to explore further security and cryptography approaches toward a secure e-book management system
Classification of Skin Cancer Images Using Convolutional Neural Network with ResNet50 Pre-trained Model
Get PDFThe skin, an astonishingly expansive organ within the human body, plays a pivotal role in safeguarding us against the environment's harsh elements. It acts as a formidable barrier, shielding our delicate internal systems from the scorching heat of the sun and the harmful effects of relentless exposure to light. Nevertheless, it is not impervious to damage, especially when subjected to excessive sunlight and the potentially hazardous ultraviolet (UV) radiation that accompanies it. Prolonged UV exposure can wreak havoc on our skin cells, potentially setting the stage for the development of skin cancer. This condition demands prompt and accurate diagnosis for effective treatment. To address the pressing need for swift and precise skin cancer diagnosis, cutting-edge technology has come to the fore in the form of deep learning systems. These sophisticated systems have been meticulously designed and trained to classify skin lesions autonomously with remarkable accuracy. The Convolutional Neural Network (CNN) architecture is a formidable choice for handling image classification tasks among the array of deep learning techniques. In a recent breakthrough study, a CNN-based model was meticulously constructed to explicitly classify skin lesions, leveraging the power of a pre-trained ResNet50 architectural model to augment its capabilities. This groundbreaking ResNet50 architecture was meticulously trained to classify seven distinct skin lesions, surpassing the performance of its predecessor, MobileNet. The results achieved in this endeavor are nothing short of impressive. The overall accuracy of the ResNet50 model stands at a commendable 87.42% when tasked with classifying the seven diverse classes within the dataset. Delving further into its proficiency, we find that the Top2 and Top3 accuracy rates soar to an astounding 95.52% and 97.86%, respectively, illustrating the model's exceptional precision and reliability
Application of Artificial Intelligence in Detecting SQL Injection Attacks
Get PDFSQL injection attacks rank among the most significant threats to data security. While AI and machine learning have advanced considerably, their application in cybersecurity remains relatively undeveloped. This work mainly aims to solve the IT-related challenge of insufficient knowledge bases and tools for security practitioners to monitor and mitigate SQL Injection attacks with AI/ML techniques. The study uses a mixed-methods approach to evaluate how well different AI and ML algorithms identify SQL injection attacks by combining algorithmic evaluation with empirical investigation. Datasets of well-known SQL injection attack patterns and AI/ML models intended for cybersecurity anomaly detection are among the resources underexplored; these findings show the potential for boosting detection capabilities by deploying ML and AI-based security solutions; specific algorithms have demonstrated success rates of up to 80% in detecting SQL injections. Despite this promising performance, around 75% of survey participants acknowledged a decrease in harmful content, with a similar number highlighting increased efficiency in their roles as security researchers or incident responders. Nevertheless, the tool’s adoption among cybersecurity professionals remains under 30%. This underscores a gap between the capabilities these technologies offer and their current level of adoption among professionals. This will help lay the groundwork for future work in identifying the best solutions and providing potential approaches to incorporating AI/ML into cybersecurity frameworks. The implications of this study indicate that adopting robust defenses against SQL injection and other cyber threats could increase many folds if we continue to research and implement AI ML. technologies
Analyzing The Impact of Project-Based Learning STEAM Flipped Classroom on Computer Architecture and Organization Courses in Higher Education
Get PDFThis study examines the differences in students' creative thinking skills between the project-based learning STEAM Flipped Classroom and the direct learning STEAM Flipped Classroom model by paying attention to the role of academic self-efficacy as a moderator variable. The research population consists of students in informatics engineering study programs in the Denpasar area of Bali. The study used a quasi-experimental 2 x 2 factorial design. Data collection is carried out through questionnaires and test instruments. Data analysis used descriptive and inferential statistics, as well as analysis of covariance. The results showed a significant difference in students' creative thinking skills between those who learned with the STEAM Flipped Classroom project learning model and the STEAM Flipped Classroom direct learning model. In addition, there are differences in students' creative thinking skills based on the level of academic self-efficacy. The empirical application of the STEAM Flipped Classroom project model is proven to help students generate new ideas, develop ideas, and improve their thinking skills. However, the findings suggest that flexible thinking skills must be further enhanced. Students with high self-efficacy tend to be more proactive in providing constructive ideas in project-based learning activities. This research implies that it is necessary to actively motivate and share experiential stories with students who have low self-efficacy. In the future, I suggest that universities adopt the innovative learning model of project-based learning STEAM flipped Classroom to improve the creative thinking skills of informatics engineering students at the college level