Pub Date : 2024-08-08DOI: 10.5815/ijitcs.2024.04.04
Akinul Islam Jony, Arjun Kumar Bose Arnob
An increase in cyber threats directed at interconnected devices has resulted from the proliferation of the Internet of Things (IoT), which necessitates the implementation of comprehensive defenses against evolving attack vectors. This research investigates the utilization of machine learning (ML) prediction models to identify and defend against cyber-attacks targeting IoT networks. Central emphasis is placed on the thorough examination of the CIC-IoT2023 dataset, an extensive collection comprising a wide range of Distributed Denial of Service (DDoS) assaults on diverse IoT devices. This ensures the utilization of a practical and comprehensive benchmark for assessment. This study develops and compares four distinct machine learning models Logistic Regression (LR), K-Nearest Neighbors (KNN), Decision Tree (DT), and Random Forest (RF) to determine their effectiveness in detecting and preventing cyber threats to the Internet of Things (IoT). The comprehensive assessment incorporates a wide range of performance indicators, such as F1-score, accuracy, precision, and recall. Significantly, the results emphasize the superior performance of DT and RF, demonstrating exceptional accuracy rates of 0.9919 and 0.9916, correspondingly. The models demonstrate an outstanding capability to differentiate between benign and malicious packets, as supported by their high precision, recall, and F1 scores. The precision-recall curves and confusion matrices provide additional evidence that DT and RF are strong contenders in the field of IoT intrusion detection. Additionally, KNN demonstrates a noteworthy accuracy of 0.9380. On the other hand, LR demonstrates the least accuracy with a value of 0.8275, underscoring its inherent incapability to classify threats. In conjunction with the realistic and diverse characteristics of the CIC-IoT2023 dataset, the study's empirical assessments provide invaluable knowledge for determining the most effective machine learning algorithms and fortification strategies to protect IoT infrastructures. Furthermore, this study establishes ground-breaking suggestions for subsequent inquiries, urging the examination of unsupervised learning approaches and the incorporation of deep learning models to decipher complex patterns within IoT networks. These developments have the potential to strengthen cybersecurity protocols for Internet of Things (IoT) ecosystems, reduce the impact of emergent risks, and promote robust defense systems against ever-changing cyber challenges.
{"title":"Securing the Internet of Things: Evaluating Machine Learning Algorithms for Detecting IoT Cyberattacks Using CIC-IoT2023 Dataset","authors":"Akinul Islam Jony, Arjun Kumar Bose Arnob","doi":"10.5815/ijitcs.2024.04.04","DOIUrl":"https://doi.org/10.5815/ijitcs.2024.04.04","url":null,"abstract":"An increase in cyber threats directed at interconnected devices has resulted from the proliferation of the Internet of Things (IoT), which necessitates the implementation of comprehensive defenses against evolving attack vectors. This research investigates the utilization of machine learning (ML) prediction models to identify and defend against cyber-attacks targeting IoT networks. Central emphasis is placed on the thorough examination of the CIC-IoT2023 dataset, an extensive collection comprising a wide range of Distributed Denial of Service (DDoS) assaults on diverse IoT devices. This ensures the utilization of a practical and comprehensive benchmark for assessment. This study develops and compares four distinct machine learning models Logistic Regression (LR), K-Nearest Neighbors (KNN), Decision Tree (DT), and Random Forest (RF) to determine their effectiveness in detecting and preventing cyber threats to the Internet of Things (IoT). The comprehensive assessment incorporates a wide range of performance indicators, such as F1-score, accuracy, precision, and recall. Significantly, the results emphasize the superior performance of DT and RF, demonstrating exceptional accuracy rates of 0.9919 and 0.9916, correspondingly. The models demonstrate an outstanding capability to differentiate between benign and malicious packets, as supported by their high precision, recall, and F1 scores. The precision-recall curves and confusion matrices provide additional evidence that DT and RF are strong contenders in the field of IoT intrusion detection. Additionally, KNN demonstrates a noteworthy accuracy of 0.9380. On the other hand, LR demonstrates the least accuracy with a value of 0.8275, underscoring its inherent incapability to classify threats. In conjunction with the realistic and diverse characteristics of the CIC-IoT2023 dataset, the study's empirical assessments provide invaluable knowledge for determining the most effective machine learning algorithms and fortification strategies to protect IoT infrastructures. Furthermore, this study establishes ground-breaking suggestions for subsequent inquiries, urging the examination of unsupervised learning approaches and the incorporation of deep learning models to decipher complex patterns within IoT networks. These developments have the potential to strengthen cybersecurity protocols for Internet of Things (IoT) ecosystems, reduce the impact of emergent risks, and promote robust defense systems against ever-changing cyber challenges.","PeriodicalId":130361,"journal":{"name":"International Journal of Information Technology and Computer Science","volume":"3 2","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141928861","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pursuing prey by a predator is a natural phenomenon. This is an event when a predator targets and chases prey for consuming. The motive of a predator is to catch its prey whereas the motive of a prey is to escape from the predator. Earth has many predator species with different pursuing strategies. Some of them are sneaky again some of them are bolt. But their chases fail every time. A successful hunt depends on the strategy of pursuing one. Among all the predators, the Dragonflies, also known as natural drones, are considered the best predators because of their higher rate of successful hunting. If their strategy of pursuing a prey can be extracted for analysis and make an algorithm to apply on Unmanned arial vehicles, the success rate will be increased, and it will be more efficient than that of a dragonfly. We examine the pursuing strategy of a dragonfly using LSTM to predict the speed and distance between predator and prey. Also, The Kalman filter has been used to trace the trajectory of both Predator and Prey. We found that dragonflies follow distance maintenance strategy to pursue prey and try to keep its velocity constant to maintain the safe (mean) distance. This study can lead researchers to enhance the new and exciting algorithm which can be applied on Unmanned arial vehicles (UAV).
{"title":"Mimicking Nature: Analysis of Dragonfly Pursuit Strategies Using LSTM and Kalman Filter","authors":"Mehedi Hassan Zidan, Rayhan Ahmed, Khandakar Anim Hassan Adnan, Tajkurun Zannat Mumu, Md. Mahmudur Rahman, D. Karmaker","doi":"10.5815/ijitcs.2024.04.06","DOIUrl":"https://doi.org/10.5815/ijitcs.2024.04.06","url":null,"abstract":"Pursuing prey by a predator is a natural phenomenon. This is an event when a predator targets and chases prey for consuming. The motive of a predator is to catch its prey whereas the motive of a prey is to escape from the predator. Earth has many predator species with different pursuing strategies. Some of them are sneaky again some of them are bolt. But their chases fail every time. A successful hunt depends on the strategy of pursuing one. Among all the predators, the Dragonflies, also known as natural drones, are considered the best predators because of their higher rate of successful hunting. If their strategy of pursuing a prey can be extracted for analysis and make an algorithm to apply on Unmanned arial vehicles, the success rate will be increased, and it will be more efficient than that of a dragonfly. We examine the pursuing strategy of a dragonfly using LSTM to predict the speed and distance between predator and prey. Also, The Kalman filter has been used to trace the trajectory of both Predator and Prey. We found that dragonflies follow distance maintenance strategy to pursue prey and try to keep its velocity constant to maintain the safe (mean) distance. This study can lead researchers to enhance the new and exciting algorithm which can be applied on Unmanned arial vehicles (UAV).","PeriodicalId":130361,"journal":{"name":"International Journal of Information Technology and Computer Science","volume":"45 21","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141928282","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-08DOI: 10.5815/ijitcs.2024.04.07
Anatoliy Litvinov, D. Chumachenko, Nataliia Dotsenko, I. Kadykova, I. Chumachenko
We address the challenge of optimizing the interaction between medical personnel and treatment stations within mobile and flexible medical care units (MFMCUs) in conflict zones. For the analysis of such systems, a closed queuing model with a finite number of treatment stations has been developed, which accounts for the possibility of performing multiple tasks for a single medical service request. Under the assumption of Poisson event flows, a system of integro-differential equations for the probability densities of the introduced states has been compiled. To solve it, the method of discrete binomial transformations is employed in conjunction with production functions. Solutions were obtained in the form of finite expressions, enabling the transition from the probabilistic characteristics of the model to the main performance metrics of the MFMCU: the load factor of medical personnel, and the utilization rate of treatment stations. The results show the selection of the number of treatment stations in the medical care area and the calculation of the appropriate performance of medical personnel.
{"title":"Enhancing Healthcare Provision in Conflict Zones: Queuing System Models for Mobile and Flexible Medical Care Units with a Limited Number of Treatment Stations","authors":"Anatoliy Litvinov, D. Chumachenko, Nataliia Dotsenko, I. Kadykova, I. Chumachenko","doi":"10.5815/ijitcs.2024.04.07","DOIUrl":"https://doi.org/10.5815/ijitcs.2024.04.07","url":null,"abstract":"We address the challenge of optimizing the interaction between medical personnel and treatment stations within mobile and flexible medical care units (MFMCUs) in conflict zones. For the analysis of such systems, a closed queuing model with a finite number of treatment stations has been developed, which accounts for the possibility of performing multiple tasks for a single medical service request. Under the assumption of Poisson event flows, a system of integro-differential equations for the probability densities of the introduced states has been compiled. To solve it, the method of discrete binomial transformations is employed in conjunction with production functions. Solutions were obtained in the form of finite expressions, enabling the transition from the probabilistic characteristics of the model to the main performance metrics of the MFMCU: the load factor of medical personnel, and the utilization rate of treatment stations. The results show the selection of the number of treatment stations in the medical care area and the calculation of the appropriate performance of medical personnel.","PeriodicalId":130361,"journal":{"name":"International Journal of Information Technology and Computer Science","volume":"12 3","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141926049","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-08DOI: 10.5815/ijitcs.2024.04.02
Md. Amdad Hossain, Mahfuzulhoq Chowdhury
The inaccurate detection of diabetes and hypertension causes’ time wastage and a cost burden due to higher amounts of medicine intake and health problems. The previous works did not investigate machine learning (ML)-based diabetic and hypertension patient prediction by using multiple characteristics. This paper utilizes ML algorithms to predict the presence of diabetes and hypertension in patients. By analyzing patient data, including medical records, symptoms, and risk factors, the proposed system can provide accurate predictions for early detection and intervention. This paper makes a list of eighteen characteristics that can be used for data set preparation. With a classification accuracy of 93%, the Support Vector Machine is the best ML model in our work and is used for the diabetic and hypertension disease prediction models. This paper also gives a new mobile application that alleviates the time and cost burden by detecting diabetic and hypertensive patients, doctors, and medical information. The user evaluation and rating analysis results showed that more than sixty five percent of users declared the necessity of the proposed application features.
糖尿病和高血压的检测不准确会造成时间浪费,并因更多的药物摄入和健康问题而造成成本负担。以往的研究没有研究基于机器学习(ML)的糖尿病和高血压患者预测,而是利用多种特征进行预测。本文利用 ML 算法预测患者是否患有糖尿病和高血压。通过分析患者数据(包括病历、症状和风险因素),所提出的系统可为早期检测和干预提供准确预测。本文列出了可用于数据集准备的十八种特征。支持向量机的分类准确率高达 93%,是我们的工作中最好的 ML 模型,并被用于糖尿病和高血压疾病预测模型。本文还给出了一个新的移动应用程序,通过检测糖尿病和高血压患者、医生和医疗信息,减轻了时间和成本负担。用户评价和评分分析结果显示,超过百分之六十五的用户表示有必要使用所提出的应用功能。
{"title":"A Machine Learning Based Intelligent Diabetic and Hypertensive Patient Prediction Scheme and A Mobile Application for Patients Assistance","authors":"Md. Amdad Hossain, Mahfuzulhoq Chowdhury","doi":"10.5815/ijitcs.2024.04.02","DOIUrl":"https://doi.org/10.5815/ijitcs.2024.04.02","url":null,"abstract":"The inaccurate detection of diabetes and hypertension causes’ time wastage and a cost burden due to higher amounts of medicine intake and health problems. The previous works did not investigate machine learning (ML)-based diabetic and hypertension patient prediction by using multiple characteristics. This paper utilizes ML algorithms to predict the presence of diabetes and hypertension in patients. By analyzing patient data, including medical records, symptoms, and risk factors, the proposed system can provide accurate predictions for early detection and intervention. This paper makes a list of eighteen characteristics that can be used for data set preparation. With a classification accuracy of 93%, the Support Vector Machine is the best ML model in our work and is used for the diabetic and hypertension disease prediction models. This paper also gives a new mobile application that alleviates the time and cost burden by detecting diabetic and hypertensive patients, doctors, and medical information. The user evaluation and rating analysis results showed that more than sixty five percent of users declared the necessity of the proposed application features.","PeriodicalId":130361,"journal":{"name":"International Journal of Information Technology and Computer Science","volume":"16 15","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141927782","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-06-08DOI: 10.5815/ijitcs.2024.03.03
Bahubali Akiwate, Sanjay Ankali, S.G. Gollagi, N. Ghani
The Internet of Things (IoT) allows you to connect a broad spectrum of smart devices through the Internet. Incorporating IoT sensors for remote health monitoring is a game-changer for the medical industry, especially in limited spaces. Environmental sensors can be installed in small rooms to monitor an individual's health. Through low-cost sensors, as the core of the IoT physical layer, the RF (Radio Frequency) identification technique is advanced enough to facilitate personal healthcare. Recently, RFID technology has been utilized in the healthcare sector to enhance accurate data collection through various software systems. Steganography is a method that makes user data more secure than it has ever been before. The necessity of upholding secrecy in the widely used healthcare system will be covered in this solution. Health monitoring sensors are a crucial tool for analyzing real-time data and developing the medical box, an innovative solution that provides patients with access to medical assistance. By monitoring patients remotely, healthcare professionals can provide prompt medical attention whenever needed while ensuring patients' privacy and personal information are protected.
{"title":"Information Security based on IoT for e-Health Care Using RFID Technology and Steganography","authors":"Bahubali Akiwate, Sanjay Ankali, S.G. Gollagi, N. Ghani","doi":"10.5815/ijitcs.2024.03.03","DOIUrl":"https://doi.org/10.5815/ijitcs.2024.03.03","url":null,"abstract":"The Internet of Things (IoT) allows you to connect a broad spectrum of smart devices through the Internet. Incorporating IoT sensors for remote health monitoring is a game-changer for the medical industry, especially in limited spaces. Environmental sensors can be installed in small rooms to monitor an individual's health. Through low-cost sensors, as the core of the IoT physical layer, the RF (Radio Frequency) identification technique is advanced enough to facilitate personal healthcare. Recently, RFID technology has been utilized in the healthcare sector to enhance accurate data collection through various software systems. Steganography is a method that makes user data more secure than it has ever been before. The necessity of upholding secrecy in the widely used healthcare system will be covered in this solution. Health monitoring sensors are a crucial tool for analyzing real-time data and developing the medical box, an innovative solution that provides patients with access to medical assistance. By monitoring patients remotely, healthcare professionals can provide prompt medical attention whenever needed while ensuring patients' privacy and personal information are protected.","PeriodicalId":130361,"journal":{"name":"International Journal of Information Technology and Computer Science","volume":" 11","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141368623","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Retinal diseases are a significant challenge in the realm of medical diagnosis, with potential complications to vision and overall ocular health. This research endeavors to address the challenge of automating the detection of retinal disease states using advanced deep learning models, including VGG-19, ResNet-50, InceptionV3, and EfficientNetV2. Each model leverages transfer learning, drawing insights from a substantial dataset comprising optical coherence tomography (OCT) images and subsequently classifying images into four distinct retinal conditions: choroidal neovascularization, drusen, diabetic macular edema and a healthy state. The training dataset, sourced from repositories that are available to the public including OCT retinal images, spanning all four disease categories. Our findings reveal that among the models tested, EfficientNetV2 demonstrates superior performance, with a remarkable classification accuracy of 98.92%, precision of 99.6%, and a recall of 99.4%, surpassing the performance of the other models.
视网膜疾病是医学诊断领域的一大挑战,对视力和整体眼部健康具有潜在的并发症。这项研究致力于利用先进的深度学习模型(包括 VGG-19、ResNet-50、InceptionV3 和 EfficientNetV2)来应对视网膜疾病状态自动检测的挑战。每个模型都利用迁移学习,从包含光学相干断层扫描(OCT)图像的大量数据集中汲取洞察力,随后将图像分类为四种不同的视网膜状况:脉络膜新生血管、色素沉着、糖尿病性黄斑水肿和健康状态。训练数据集来自公共资源库,包括 OCT 视网膜图像,涵盖所有四种疾病类别。我们的研究结果表明,在所测试的模型中,EfficientNetV2 表现出卓越的性能,其分类准确率高达 98.92%,精确率高达 99.6%,召回率高达 99.4%,超过了其他模型。
{"title":"Advanced Deep Learning Models for Accurate Retinal Disease State Detection","authors":"Hossein. Abbasi, Ahmed. Alshaeeb, Yasin Orouskhani, Behrouz. Rahimi, Mostafa Shomalzadeh","doi":"10.5815/ijitcs.2024.03.06","DOIUrl":"https://doi.org/10.5815/ijitcs.2024.03.06","url":null,"abstract":"Retinal diseases are a significant challenge in the realm of medical diagnosis, with potential complications to vision and overall ocular health. This research endeavors to address the challenge of automating the detection of retinal disease states using advanced deep learning models, including VGG-19, ResNet-50, InceptionV3, and EfficientNetV2. Each model leverages transfer learning, drawing insights from a substantial dataset comprising optical coherence tomography (OCT) images and subsequently classifying images into four distinct retinal conditions: choroidal neovascularization, drusen, diabetic macular edema and a healthy state. The training dataset, sourced from repositories that are available to the public including OCT retinal images, spanning all four disease categories. Our findings reveal that among the models tested, EfficientNetV2 demonstrates superior performance, with a remarkable classification accuracy of 98.92%, precision of 99.6%, and a recall of 99.4%, surpassing the performance of the other models.","PeriodicalId":130361,"journal":{"name":"International Journal of Information Technology and Computer Science","volume":" 13","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141368949","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This research suggests an efficient idea that is better suited for speech processing applications for retrieving the accurate pitch from speech signal in noisy conditions. For this objective, we present a fundamental frequency extraction algorithm and that is tolerant to the non-stationary changes of the amplitude and frequency of the input signal. Moreover, we use an accumulated power spectrum instead of power spectrum, which uses the shorter sub-frames of the input signal to reduce the noise characteristics of the speech signals. To increase the accuracy of the fundamental frequency extraction we have concentrated on maintaining the speech harmonics in their original state and suppressing the noise elements involved in the noisy speech signal. The two stages that make up the suggested fundamental frequency extraction approach are producing the accumulated power spectrum of the speech signal and weighting it with the average magnitude difference function. As per the experiment results, the proposed technique appears to be better in noisy situations than other existing state-of-the-art methods such as Weighted Autocorrelation Function (WAF), PEFAC, and BaNa.
这项研究提出了一种更适合语音处理应用的高效思路,可在噪声条件下从语音信号中提取准确的音高。为此,我们提出了一种基频提取算法,该算法对输入信号的振幅和频率的非稳态变化具有容忍性。此外,我们使用累积功率谱代替功率谱,利用输入信号的较短子帧来降低语音信号的噪声特性。为了提高基频提取的准确性,我们集中精力保持语音谐波的原始状态,并抑制噪声语音信号中的噪声元素。所建议的基频提取方法分为两个阶段,一是生成语音信号的累积功率谱,二是用平均幅度差函数对其进行加权。实验结果表明,与加权自相关函数 (WAF)、PEFAC 和 BaNa 等其他现有的先进方法相比,建议的技术在噪声环境中的效果更好。
{"title":"Fundamental Frequency Extraction by Utilizing Accumulated Power Spectrum based Weighted Autocorrelation Function in Noisy Speech","authors":"Nargis Parvin, Moinur Rahman, Irana Tabassum Ananna, Md. Saifur Rahman","doi":"10.5815/ijitcs.2024.03.05","DOIUrl":"https://doi.org/10.5815/ijitcs.2024.03.05","url":null,"abstract":"This research suggests an efficient idea that is better suited for speech processing applications for retrieving the accurate pitch from speech signal in noisy conditions. For this objective, we present a fundamental frequency extraction algorithm and that is tolerant to the non-stationary changes of the amplitude and frequency of the input signal. Moreover, we use an accumulated power spectrum instead of power spectrum, which uses the shorter sub-frames of the input signal to reduce the noise characteristics of the speech signals. To increase the accuracy of the fundamental frequency extraction we have concentrated on maintaining the speech harmonics in their original state and suppressing the noise elements involved in the noisy speech signal. The two stages that make up the suggested fundamental frequency extraction approach are producing the accumulated power spectrum of the speech signal and weighting it with the average magnitude difference function. As per the experiment results, the proposed technique appears to be better in noisy situations than other existing state-of-the-art methods such as Weighted Autocorrelation Function (WAF), PEFAC, and BaNa.","PeriodicalId":130361,"journal":{"name":"International Journal of Information Technology and Computer Science","volume":" 45","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141370228","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The objective of this PRISMA-Driven systematic review is to analyze the relative progress of Urdu speech recognition for the very first time by comparing it mainly with three selected languages; English, Mandarin Chinese, and Hindi based on Artificially Intelligent (AI) building blocks i.e. datasets, feature extraction techniques, experimental design, acoustic and language models. The selection of languages embarks from the speakers of a particular language which reveals that the chosen languages are the world's top spoken languages while Urdu ranks at number ten and is continuously progressing. A total of 176 articles were extracted from the Google Scholar database using custom queries for each language. Among them, 47 articles were selected including 5 review articles and 42 research articles, as per our inclusion criteria and after undergoing quality assessment checks. Comparative research has been designed and findings were organized based on four possible speech types i.e. spontaneous, continuous, connected words and isolated words; twenty-one datasets inclusive benchmark; MFCC, Triangular, Mel spectrogram and Log Mel features; state-of-the-art acoustic and language models; and recognition performance. The findings presented in this systematic literature review have enlightened Urdu and Hindi research towards the best available AI and deep learning practices of English and Mandarin Chinese primarily Triangular filters, Mel spectrogram, Transformers, and Attention as these techniques reveal recent trends and achieved breakthrough performance evident by their word error rate, character error rate, and perplexity.
{"title":"A PRISMA-driven Review of Speech Recognition based on English, Mandarin Chinese, Hindi and Urdu Language","authors":"Muhammad Hazique Khatri, Humera Tariq, Maryam Feroze, Ebad Ali, Zeeshan Anjum Junaidi","doi":"10.5815/ijitcs.2024.03.04","DOIUrl":"https://doi.org/10.5815/ijitcs.2024.03.04","url":null,"abstract":"The objective of this PRISMA-Driven systematic review is to analyze the relative progress of Urdu speech recognition for the very first time by comparing it mainly with three selected languages; English, Mandarin Chinese, and Hindi based on Artificially Intelligent (AI) building blocks i.e. datasets, feature extraction techniques, experimental design, acoustic and language models. The selection of languages embarks from the speakers of a particular language which reveals that the chosen languages are the world's top spoken languages while Urdu ranks at number ten and is continuously progressing. A total of 176 articles were extracted from the Google Scholar database using custom queries for each language. Among them, 47 articles were selected including 5 review articles and 42 research articles, as per our inclusion criteria and after undergoing quality assessment checks. Comparative research has been designed and findings were organized based on four possible speech types i.e. spontaneous, continuous, connected words and isolated words; twenty-one datasets inclusive benchmark; MFCC, Triangular, Mel spectrogram and Log Mel features; state-of-the-art acoustic and language models; and recognition performance. The findings presented in this systematic literature review have enlightened Urdu and Hindi research towards the best available AI and deep learning practices of English and Mandarin Chinese primarily Triangular filters, Mel spectrogram, Transformers, and Attention as these techniques reveal recent trends and achieved breakthrough performance evident by their word error rate, character error rate, and perplexity.","PeriodicalId":130361,"journal":{"name":"International Journal of Information Technology and Computer Science","volume":" 29","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141369988","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-06-08DOI: 10.5815/ijitcs.2024.03.02
Shourove Sutradhar Dip, Md. Habibur Rahman, Nazrul Islam, Md. Easin Arafat, Pulak Kanti Bhowmick, Mohammad Abu Yousuf
Brain tumors are among the deadliest forms of cancer, and there is a significant death rate in patients. Identifying and classifying brain tumors are critical steps in understanding their functioning. The best way to treat a brain tumor depends on its type, size, and location. In the modern era, Radiologists utilize Brain tumor locations that can be determined using magnetic resonance imaging (MRI). However, manual tests and MRI examinations are time-consuming and require skills. In addition, misdiagnosis of tumors can lead to inappropriate medical therapy, which could reduce their chances of living. As technology advances in Deep Learning (DL), Computer Assisted Diagnosis (CAD) as well as Machine Learning (ML) technique has been developed to aid in the detection of brain tumors, radiologists can now more accurately identify brain tumors. This paper proposes an MRI image classification using a VGG16 model to make a deep convolutional neural network (DCNN) architecture. The proposed model was evaluated with two sets of brain MRI data from Kaggle. Considering both datasets during the training at Google Colab, the proposed method achieved significant performance with a maximum overall accuracy of 96.67% and 97.67%, respectively. The proposed model was reported to have worked well during the training period and been highly accurate. The proposed model's performance criteria go beyond existing techniques.
{"title":"Enhancing Brain Tumor Classification in MRI: Leveraging Deep Convolutional Neural Networks for Improved Accuracy","authors":"Shourove Sutradhar Dip, Md. Habibur Rahman, Nazrul Islam, Md. Easin Arafat, Pulak Kanti Bhowmick, Mohammad Abu Yousuf","doi":"10.5815/ijitcs.2024.03.02","DOIUrl":"https://doi.org/10.5815/ijitcs.2024.03.02","url":null,"abstract":"Brain tumors are among the deadliest forms of cancer, and there is a significant death rate in patients. Identifying and classifying brain tumors are critical steps in understanding their functioning. The best way to treat a brain tumor depends on its type, size, and location. In the modern era, Radiologists utilize Brain tumor locations that can be determined using magnetic resonance imaging (MRI). However, manual tests and MRI examinations are time-consuming and require skills. In addition, misdiagnosis of tumors can lead to inappropriate medical therapy, which could reduce their chances of living. As technology advances in Deep Learning (DL), Computer Assisted Diagnosis (CAD) as well as Machine Learning (ML) technique has been developed to aid in the detection of brain tumors, radiologists can now more accurately identify brain tumors. This paper proposes an MRI image classification using a VGG16 model to make a deep convolutional neural network (DCNN) architecture. The proposed model was evaluated with two sets of brain MRI data from Kaggle. Considering both datasets during the training at Google Colab, the proposed method achieved significant performance with a maximum overall accuracy of 96.67% and 97.67%, respectively. The proposed model was reported to have worked well during the training period and been highly accurate. The proposed model's performance criteria go beyond existing techniques.","PeriodicalId":130361,"journal":{"name":"International Journal of Information Technology and Computer Science","volume":" 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141368273","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-06-08DOI: 10.5815/ijitcs.2024.03.01
C. P. Olipas, Ruth G. Luciano
This study presents a comprehensive assessment of the test performance of Bachelor of Science in Information Technology (BSIT) students in the System Integration and Architecture (SIA) course, coupled with a meticulous examination of the quality of test questions, aiming to lay the groundwork for enhancing the assessment tool. Employing a cross-sectional research design, the study involved 200 fourth-year students enrolled in the course. The results illuminated a significant discrepancy in scores between upper and lower student cohorts, highlighting the necessity for targeted interventions, curriculum enhancements, and assessment refinements, particularly for those in the lower-performing group. Further examination of the item difficulty index of the assessment tool unveiled the need to fine-tune certain items to better suit a broader spectrum of students. Nevertheless, the majority of items were deemed adequately aligned with their respective difficulty levels. Additionally, an analysis of the item discrimination index identified 25 items suitable for retention, while 27 items warranted revision, and 3 items were suitable for removal, as per the analysis outcomes. These insights provide a valuable foundation for improving the assessment tool, thereby optimizing its capacity to evaluate students' acquired knowledge effectively. The study's novel contribution lies in its integration of both student performance assessment and evaluation of assessment tool quality within the BSIT program, offering actionable insights for improving educational outcomes. By identifying challenges faced by BSIT students and proposing targeted interventions, curriculum enhancements, and assessment refinements, the research advances our understanding of effective assessment practices. Furthermore, the detailed analysis of item difficulty and discrimination indices offers practical guidance for enhancing the reliability and validity of assessment tools in the BSIT program. Overall, this research contributes to the existing body of knowledge by providing empirical evidence and actionable recommendations tailored to the needs of BSIT students, promoting educational quality and student success in Information Technology.
{"title":"Analyzing Test Performance of BSIT Students and Question Quality: A Study on Item Difficulty Index and Item Discrimination Index for Test Question Improvement","authors":"C. P. Olipas, Ruth G. Luciano","doi":"10.5815/ijitcs.2024.03.01","DOIUrl":"https://doi.org/10.5815/ijitcs.2024.03.01","url":null,"abstract":"This study presents a comprehensive assessment of the test performance of Bachelor of Science in Information Technology (BSIT) students in the System Integration and Architecture (SIA) course, coupled with a meticulous examination of the quality of test questions, aiming to lay the groundwork for enhancing the assessment tool. Employing a cross-sectional research design, the study involved 200 fourth-year students enrolled in the course. The results illuminated a significant discrepancy in scores between upper and lower student cohorts, highlighting the necessity for targeted interventions, curriculum enhancements, and assessment refinements, particularly for those in the lower-performing group. Further examination of the item difficulty index of the assessment tool unveiled the need to fine-tune certain items to better suit a broader spectrum of students. Nevertheless, the majority of items were deemed adequately aligned with their respective difficulty levels. Additionally, an analysis of the item discrimination index identified 25 items suitable for retention, while 27 items warranted revision, and 3 items were suitable for removal, as per the analysis outcomes. These insights provide a valuable foundation for improving the assessment tool, thereby optimizing its capacity to evaluate students' acquired knowledge effectively. The study's novel contribution lies in its integration of both student performance assessment and evaluation of assessment tool quality within the BSIT program, offering actionable insights for improving educational outcomes. By identifying challenges faced by BSIT students and proposing targeted interventions, curriculum enhancements, and assessment refinements, the research advances our understanding of effective assessment practices. Furthermore, the detailed analysis of item difficulty and discrimination indices offers practical guidance for enhancing the reliability and validity of assessment tools in the BSIT program. Overall, this research contributes to the existing body of knowledge by providing empirical evidence and actionable recommendations tailored to the needs of BSIT students, promoting educational quality and student success in Information Technology.","PeriodicalId":130361,"journal":{"name":"International Journal of Information Technology and Computer Science","volume":" 18","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141368163","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}