The Internet is considered the greatest invention of human history, as it serves as the collection of information of mankind. However, given the internet’s distributed nature, fortunately and unfortunately there are little to no regulations on information dissemination. With the internet becoming more and more accessible as infrastructure becomes better and Internet Service Providers offer cheaper services, there has been an increase in people possessing the ability to share and access information. This virtue, however, is plagued by false information present on the internet backed by various motivations ranging from malicious intentions to satire. A big part of this false information consists of inaccurate content presented as news often to change public opinion this is also called ‘fake news’. Digital platforms such as Facebook, Twitter, etc. have a major contribution towards the generation, and propagation of fake news, while such information can be reported and taken down from the platforms, current measures taken come off as bleak and need more work. Detection of false news can lead to significant impact on the masses that majorly lack media literacy. To address this issue the paper proposes combining linguistic feature extraction and word embeddings to create feature vectors that will improve classification of fake news, also the methodologies are tested on the WELfake dataset.

Medical disease datasets are heterogeneous in nature having varying dimensions and imbalanced features, which makes learning models more difficult in predicting the disease classes effectively. Traditionally, machine learning (ML) models have been successfully employed in several studies to improve illness prediction models efficiently. In this research, an enhanced feature extraction and fine-tuned ensemble learning model is proposed which combines bagging, boosting and weak learner's characteristics to effectively enhance the performance of the model. The experimental assessments are then performed on four distinct disease datasets of Parkinson's, diabetes, breast cancer and cardiovascular disorders. The datasets are first pre-processed by eliminating rows with null values and then they are standardized as per the requirement. Feature extraction is given significant importance by applying two statistical approaches such as chi-square and Info Gain to reduce the multicolinearity problem among four different datasets for multi disease classification. The suggested model has been assessed using a range of various unique performance indicators such as precision, accuracy, recall, F1score, hamming loss, log loss, Cohen Kappa score, Jaccard score, Matthew's correlation coefficient, and misclassification mistake. Experiments have validated the effectiveness of the suggested ensemble learning paradigm compared to the state-of-the-art ML and other models. The ensemble learning with enhanced feature extraction model has achieved accuracy 98.6% for breast cancer, 83.15% for cardiovascular disorder, 91.69% for parkinson's disease and 92.18% for diabetes prediction.

COVID-19 has had a profound impact on global health, targeting the human respiratory system and causing significant disruptions to human life. To aid in effective diagnosis and classification, numerous machine learning and deep learning models have been developed to analyse limited Chest Radiographic and Computed Tomography images. In this study, we propose a three-layer stacked multimodal approach for deep feature extraction from a high volume of COVID-19 Chest Radiographic images. The proposed model utilizes eight transfer learning models pre-trained on the ImageNet dataset, evaluated based on key performance metrics such as accuracy, precision, and recall. The unique stacking model integrates the outputs of these transfer learning models, extracting deep features through a three-layer architecture. These features are flattened, concatenated, and passed through seven deep dense layers with varying kernel and bias dimensions to achieve optimal classification performance. The approach was applied to a large COVID-19 Chest Radiographic images dataset, achieving the highest accuracy (95.79%), precision (95.44%), and recall (96.65%) when compared to state-of-the-art models. This study demonstrates the effectiveness of leveraging a stacked transfer learning multimodal framework for COVID-19 diagnosis. The proposed method not only ensures high accuracy but also provides a computationally efficient solution for analysing large-scale radiographic datasets, positioning it as a robust tool for aiding in the early detection and classification of COVID-19.

In the digital world of remarkable technological advancements, the detection of cervical cancer at early stages is of important clinical significance as it can vastly improve the survival rate during treatment. Cervical cytopathology, often known as a Pap test is the frequently adopted screening method for cervical cancer. However, the test seems to be effective but investigation of images containing Pap smear with the help of a microscope is a difficult as well as laborious exercise. The procedure for the same demands an expert in the area and is often time-consuming. The serious pitfalls in subjective clinical evaluation evoke the need of developing an automated system for more reliable cervical cancer diagnosis. Therefore, the goal of this study primarily focuses on designing a Deep learning model to process the Pap smear images and correctly classify the cervical cells. For this purpose, firstly, a publically available dataset namely SIPaKMeD is utilized. Then, different data pre-processing methods are applied to intensify the data quality for effective analysis. Next, a novel stacking model is proposed that leverages a Support Vector Classifier (SVC) as a Meta model over a combination of different Transfer Learning Models including VGG16, ResNet101, InceptionV3, Xception, DenseNet169, and Inception ResNet. Furthermore, the dense layers are added to tune the underlying base transfer learning models to learn fine-tuned adaptive weights. The results obtained from experimental evaluation demonstrate the efficacy of the proposed stacking model by yielding the highest accuracy rate of 95.66% in comparison to other employed methods and existing state-of-the-art techniques.

Parkinson’s disease is a brain disorder that causes unintended or uncontrollable movements, such as shaking, stiffness, and difficulty with balance and coordination.As a result, there is a necessity to create an effective method capable of detecting Parkinson’s disease. The data used in this study was taken from publicly available database. Three feature selection techniques such as Chi Square, Information Gain and Grey Wolf were used for the process of choosing the best subset of features for the target data.In this study a Stacking Master Model is proposed and the basic idea behind this approach is when two or more learning techniques work in a collaborative manner will enhance the learning of the resultant model. In the proposed work some state-of-the-art learning algorithms and a proposed master model using stacking were compared to evaluate the performance of these learning algorithms. Following the application of all of these algorithms, it is discovered that the proposed master model Stacking combined with Grey Wolf produces the best results, with an accuracy of 94.17

Radio-Frequency Identification Technology (RFID) is a non-contact identification technology based on radio waves. It is mainly comprised of an RFID reader and RFID tags where RFID tags store unique information while an RFID reader emits radio waves to interrogate the tag. RFID technology doesn’t require a line of sight, so it can identify objects from a larger distance, and it has numerous applications in different areas like IoT (Internet of Things), IoV (Internet of Vehicles), etc. RFID technology relies on wireless communication for data transmission, so it is more prone to security attacks, and if security is compromised, then it may lead to disastrous results. To maintain security in RFID systems, there is a need to use some lightweight authentication protocols, as RFID tags have limited memory and less computation capabilities. This paper reviewed the three RFID-based lightweight authentication protocols, i.e. LRSAS + , LRARP, and LRARP + . Also, we have done the cryptanalysis of these lightweight protocols and found that they are unsuitable for low-cost tags due to their high computation power. So, keeping the focus on low-cost tags, we have proposed an ultra-lightweight RFID authentication protocol known as ULRARP + protocol that solely relies on ultra-lightweight operations like XOR, Perm and Rotation operations. Informal security evaluation indicates that the proposed protocol resists different security attacks. Also, a formal evaluation of the proposed protocol is done using GNY logic. Further, performance evaluation of the ULRARP + protocol is computed based on three parameters, i.e. Tag's computational cost, Storage space, and the message transmitted by the tag. Results indicate that our proposed ultra-lightweight protocol is more secure and performance-efficient than the existing state-of-the-art authentication protocols.

This paper introduces FeaTrim-ViT, a novel framework for trimming Vision Transformer (ViT) models, which are known for their exceptional performance but also for their large number of parameters, making them challenging to deploy on memory-limited devices. Our approach innovatively combines structured pruning techniques with nature-inspired algorithms to efficiently identify and eliminate redundant parameters within ViT models. Existing studies fail to maintain accuracy at higher compression rates during efficient finetuning. The framework is designed to comprehensively search for an optimal submodel, targeting multiple dimensions including input tokens, Multi-Head Self-Attention (MHSA), and Multilayer Perceptron (MLP) modules. A learnable sparsity constraint with preset factors that reflect global importance across these dimensions. The search process is based on a single-shot training scheme making it exceptionally efficient. The resulting substructures boast diverse dimensionalities, tailored to specific module requirements. A budget threshold, based on the desired accuracy-FLOPs trade-off, guides the retraining process to produce the optimised model which finally goes through efficient feature selection using Grey Wolf Optimiser (GWO). The extensive experiments demonstrate that the framework can reduce up to 35

Nowadays, wireless technology has been widely used in healthcare and communication systems. It makes our life easier in all respects. A radiofrequency identification device (RFID) has been deployed as a wireless and identity communication device. RFID is a low-resource device that requires cryptography with limited energy regarding the minimum key size. Security and authentication between the server and the tags are key challenges for an RFID system to maintain data privacy. This article presents the security vulnerabilities of recent existing RFID authentication schemes. Keeping the focus on stringent security, privacy, and low cost, we have designed a new lightweight group authentication protocol for the robust RFID system. A single server controls multiple tags by using the proposed lightweight protocol in the RFID system. Formal and informal security analysis is performed compared to other lightweight group authentication articles in which only informal security analysis is carried out. The formal security strength of our proposed protocol is analyzed using the AVISPA (Automated Verification of Internet Security Protocol Analysis) tool, confirming that it is safe from different security threats. The newly designed protocol’s performance analysis results are measured in terms of computational cost, storage space, and communication cost. Finally, the combined consequence of security and lightweight of the proposed group authentication protocol is superior and outperforms compared to the existing scheme.

Predicting cardiac or heart disease has emerged as a formidable challenge in the medical domain recently. It is recognized as a major global health concern, and stands as one of the primary causes of mortality, posing a significant threat to human life. Early detection of heart disease helps to reduce mortality. This study has experimented with three benchmark datasets such as UCI Heart Disease, Framingham, and Z-Alizadeh Saini containing important clinical information for cardiac vascular disease (CVD). These three datasets’ multi-variant (categorical and continuous) features, variable dimensions, and multicollinearity characteristics provide substantial challenges for machine learning (ML) and other models aiming to achieve the desired results. This study proposes a statistical feature selection (SFS) stacking framework using four feature engineering techniques, Chi-Square, Gini Index, Information Gain, and ANOVA F-test, to select the optimal features from the datasets. Further, the likelihood of developing CVD based on characteristics extracted from the three benchmark datasets using a reduced set of optimized features from the initial feature set is fed to ensemble stacked learning models: stacking using Support Vector Machine (SFS-SVM) and stacking using Cross-Validation Classifier (SFS-SCVC). The SFS-SCVC model has achieved significant performance metrics and outperformed the SFS-SVM and traditional ML models on all three datasets.

Skin cancers, especially melanomas and carcinomas, are among the most fatal diseases affecting the integumentary system (skin). Effective diagnosis of these ailments at an early stage opens the door to proper care, effective treatment plans, and higher life expectancy. However, diagnosis of malignancy in dermatology is a time-consuming process and is subject to a lot of human errors because of striking visual resemblance among distinct categories of lesions. Therefore, computational methods are being employed nowadays to increase the accuracy of diagnostic processes. An automated AI-based detection system can help dermatologists worldwide diagnose and classify skin lesions at the earliest possible time. In this study, we introduce a technique for classifying skin lesions that is built on an ensemble of eight deep learning architectures. The classifiers VGG-16, VGG-17, ResNet50, InceptionV3, InceptionRestnetV2, Xception, DenseNet121, and MobileNetV2 were developed using the training dataset along with fine-tuning of the parameters. These models were later integrated using the stacking approach and combination rule of majority voting. Test accuracy and test loss of 85.26

Coronary Heart Disease (CHD) is a significant public health concern, posing a substantial burden on individuals, families and healthcare systems, necessitating accurate prediction and timely diagnosis for effective management. This research paper presents a data-driven approach for enhancing CHD prediction using meticulous preprocessing to address null values, ensuring data integrity, dataset resampling to address imbalance and Min-Max normalization which further enhances the comparability of features for the Framingham CHD dataset. A Stacking Ensemble Model with Random Forest as the meta classifier is proposed, that combines the predictive capabilities of the base classifiers including Logistic Regression, KNN, SVM, Decision Tree and XGBoost. The proposed approach yields exceptional results with an accuracy of 97.39

Breast cancer (BC) is one of the leading causes of death among women worldwide, as it has emerged as the most commonly diagnosed malignancy in women.Early detection and effective treatment of BC can help save women's lives.Developing an efficient technology-based detection system can lead to non-destructive and preliminary cancer detection techniques.This paper proposes a comprehensive framework that can effectively diagnose cancerous cells from benign cells using the Curated Breast Imaging Subset of the Digital Database for Screening Mammography (CBIS-DDSM) data set.The novelty of the proposed framework lies in the integration of various techniques, where the fusion of deep learning (DL), traditional machine learning (ML) techniques, and enhanced classification models have been deployed using the curated dataset.The analysis outcome proves that the proposed enhanced RF (ERF), enhanced DT (EDT) and enhanced LR (ELR) models for BC detection outperformed most of the existing models with impressive results.

Download This Paper Open PDF in Browser Add Paper to My Library Share: Permalink Using these links will ensure access to this page indefinitely Copy URL Copy DOI

This paper explores the evolving landscape of artificial intelligence (AI) in the realm of neurodegenerative disease detection. Neurodegenerative diseases characterized by the progressive degeneration of the nervous system have significant challenges in early diagnosis and management. The conventional diagnosis was based on manual analysis of complex structures of the brain captured using neuroimaging techniques such as magnetic resonance imaging (structural MRI (sMRI), functional MRI (fMRI), resting-state MRI (rs-MRI)), diffusion-weighted (DW)-MR imaging, computed tomography (CT), single-photon emission computed tomography (SPECT), and positron emission tomography (PET). In recent times, various patterns recognition methods have been determined to discriminate automatically among the patients with and without disease of Alzheimer. Recent advancements in AI techniques, including machine learning (ML), deep learning (DL), and computer vision, have shown great promise in enhancing our ability to detect these diseases at earlier stages and with greater accuracy. The paper discusses key AI-driven methodologies, such as the analysis of neuroimaging data, the utilization of wearable devices for continuous monitoring, and the integration of multimodal data sources, which have revolutionized the field of neurodegeneration detection. Additionally, it delves into the ethical considerations and challenges associated with the widespread adoption of AI in clinical practice, emphasizing the need for data privacy, interpretability, and regulatory frameworks. Through a comprehensive review of recent literature and case studies, this paper provides insights into how AI is reshaping the landscape of neurodegenerative disease diagnosis, prognosis, and personalized treatment. It also sheds light on potential future directions and opportunities for further research and collaboration in harnessing the full potential of AI to address the global challenge of neurodegenerative diseases.

In today's world, there is a lot of anxiety about suicidal thoughts that is conveyed on social media platforms, and people are now sharing all kinds of feelings on social media forums. The majority of them utilize social forums because they feel uncomfortable sharing privately. The study's objective is to analyze suicidal thoughts and identify them at an early stage by utilizing deep learning and transfer learning techniques. These algorithms are used to data gathered from users of Reddit forums who have suicidal thoughts as well as regular users who have non-suicidal thoughts. For the aforementioned goal, we use techniques such as the transfer learning algorithms BERT, RoBERTa, and ALBERT, as well as BiLSTM and other deep learning algorithms. In contrast to the sequence processing model, our study demonstrates that the bidirectional long-short term algorithm provides the best validation accuracy, while the pretrained models BERT and ALBERT also provide satisfactory accuracy.

Diabetes is one of the fastest-growing human diseases worldwide and poses a significant threat to the population's longer lives. Early prediction of diabetes is crucial to taking precautionary steps to avoid or delay its onset. In this study, we proposed a Deep Dense Layer Neural Network (DDLNN) for diabetes prediction using a dataset with 768 instances and nine variables. We also applied a combination of classical machine learning (ML) algorithms and ensemble learning algorithms for the effective prediction of the disease. The classical ML algorithms used were Support Vector Machine (SVM), Logistic Regression (LR), Decision Tree (DT), K-Nearest Neighbor (KNN), and Naïve Bayes (NB). We also constructed ensemble models such as bagging (Random Forest) and boosting like AdaBoost and Extreme Gradient Boosting (XGBoost) to evaluate the performance of prediction models. The proposed DDLNN model and ensemble learning models were trained and tested using hyperparameter tuning and K-Fold cross-validation to determine the best parameters for predicting the disease. The combined ML models used majority voting to select the best outcomes among the models. The efficacy of the proposed and other models was evaluated for effective diabetes prediction. The investigation concluded that the proposed model, after hyperparameter tuning, outperformed other learning models with an accuracy of 84.42%, a precision of 85.12%, a recall rate of 65.40%, and a specificity of 94.11%.

Cardiovascular disease (CVD) is a formidable public health challenge across the globe and is the most prevalent cause of mortality. Early detection and accurate prediction of CVD can help prevent disease progression and reduce the risk of complications. Machine learning (ML) techniques show promising results in improving the accuracy and efficiency of CVD prediction to precision. However, the effectiveness of machine learning algorithms in CVD prediction largely depends on the selection of relevant features from complex datasets. The performance and interpretability of ML models are improved by feature selection strategies, which attempt to identify significant attributes while eliminating duplicate or irrelevant features. The feature selection and ML algorithms for CVD are thoroughly reviewed in this publication. The review provides insight into the selection of appropriate feature selection techniques and machine learning algorithms for accurate CVD prediction and evaluates the effectiveness and performance of these methods on cardiovascular datasets. Insights from the findings of this study can be used for interpreting the selection of optimal feature selection methods and ML algorithms for the precise prediction of cardiovascular disease, thereby improving patient outcomes and reducing healthcare costs.

The paper defines a new model for Takri touching text segmentation (T3S), which uses simpler statistical operations to locate exact segmentation column and generate faster results. An analysis of existing Indian scripts touching text segmentation techniques is also provided in the paper. The analysis inferred higher accuracies with recognition-based approaches which are computationally and time extensive and, work better with huge datasets. Also, most of these have been developed for headline Indian scripts only. The proposed T3S model provides significant segmentation accuracy using a statistical approach thus, generating faster results. Also, it works well with small datasets, which is feasible with the diverse ancient low-resource scripts of India like Takri. The model has been implemented on Takri text, a class of non-headline Indian regional scripts. Thus, it provides a benchmark algorithm for conducting further research in the field. A dataset of 1465 touching consonant pairs in printed Takri script is prepared using connected component segmentation for segmenting Takri text from the manually collected archival data in printed Takri script. The T3S technique is implemented on the dataset prepared and the results are critically analyzed with the existing Indian scripts touching text segmentation approaches, based on the accuracy achieved.

In present-day society, the major critical issues are mental health problems which eventually turn out to be suicidal ideation. Premature detection of suicidal thoughts among people is the solution to avoid suicide in the latter times. With each passing year, the growth rate of suicides is abruptly increasing. The social media platform is the key from which people around the world come across and share their feelings, emotions, and reaction to what they are going through in their lives. The data which is available on social media can be used for the identification and detection of suicidal ideation among people, on social media forums like Twitter, Reddit, Tumblr, Facebook, and Instagram. Natural Language Processing has come a long way in the research of finding the sentiment of individuals and checking the linguistic patterns of text shared by the people. Effective to carry out the research by using machine learning, deep learning, and transfer learning algorithms on the online forum data of people. Algorithms such as SVM, Logistic Regression, Naive Bayes, Convolution Neural Network, Recurrent Neural Network, Bidirectional Long Short-Term Memory, BERT, and RoBERTa, respectively, were enforced to classify or detect the suicides in the social media data.