Fulltext search in archive

The concept of individualization has emerged as an essential advance in education, representing a paradigm shift adopted by educational systems worldwide. This paradigm evolution aims to optimize student performance by harnessing the potential of diverse e-learning platforms tailored to individual needs. These platforms enable students to acquire knowledge that matches their unique interests and skills. This not only fosters academic prowess but also proactive engagement of stakeholders invested in talent development endeavours. This paper attempts to provide a holistic overview of adaptive e-learning approaches using a recognized categorization framework. This categorization delineates three main approaches: the macro-adaptive approach, the aptitude-treatment interaction approach and the micro-adaptive approach. Through the exposition of illustrative examples within each category, the paper seeks to elucidate and scrutinize the similarities and differences inherent in these approaches. Furthermore, it endeavours to outline the theoretical foundations of the learning process. It contrasts e-learning with traditional learning modalities and examines the transformative impact of e-learning on the learning process and knowledge generation. By delving into the historical trajectory of adaptive e-learning platforms, the paper reveals a symbiotic relationship between their evolution and the evolving principles of education and cognitive science. Ultimately, the study strives to bring a nuanced understanding of the adaptive e-learning landscape and its evolutionary trajectory over time, thereby contributing to the broader discourse surrounding educational innovation and pedagogical progress.

The use of agile methodology in software development projects is growing rapidly among industry professionals and academia. The Unified Modelling Language (UML) conventionally accompanies agile software development to model the software requirements. The user story is fundamental and should be identified to communicate the basic requirements between the development team and the stakeholders before the UML model such as the use case diagram, class diagrams and many others can be designed. However, there are several challenges associated with this process such as poorly organised user stories, natural language complexity and high time consumption to create them. A systematic literature review is conducted to grasp more knowledge about the utilisation of natural language processing (NLP) for UML model generation. A total of 198 papers were initially found in four online databases, namely Scopus, IEEE Xplore, ScienceDirect and ACM Digital Library, from the period 2018-2022. After removing duplicates, applying inclusion and exclusion criteria, and conducting the full-text assessment, only 20 papers are included as the primary studies. The primary studies are reviewed to discover several important pieces of information, namely the challenges of designing UML models, NLP tools and techniques used to generate UML models, UML models generated, and validation methods used for measuring the accuracy of generated models. Finally, this study discusses important elements related to UML model generation using NLP tools and techniques.

Person identification through chest X-ray radiographs stands as a vanguard in both healthcare and biometrical security domains. In contrast to traditional biometric modalities, such as facial recognition, fingerprints and iris scans, the research orientation towards chest X-ray recognition has been spurred by its remarkable recognition rates. Capturing the intricate anatomical nuances of an individual's rib cage, lungs and heart, chest X-ray images emerge as a focal point for identification, even in scenarios where the human body is entirely damaged. Concerning the field of deep learning, a paradigm is exemplified in contemporary generations, with promising outcomes in classification and image similarity challenges. However, the training of convolutional neural networks (CNNs) requires copious labelled data and is time-consuming. In this study, we delve into the rich repository of the NIH ChestX-ray14 dataset, comprising 112,120 frontal-view chest radiographs from 30,805 unique patients. Our methodology is nuanced, employing the potency of Siamese neural networks and the triplet loss in conjunction with refined CNN models for feature extraction. The Siamese networks facilitate robust image similarity comparison, while the triplet loss optimizes the embedding space, mitigating intra-class variations and amplifying inter-class distances. A meticulous examination of our experimental results reveals profound insights into our model performance. Noteworthy is the remarkable accuracy achieved by the VGG-19 model, standing at an impressive 97%. This achievement is underpinned by a well-balanced precision of 95.3% and an outstanding recall of 98.4%. Surpassing other CNN models utilized in our research and outshining existing state-of-the-art models, our approach establishes itself as a vanguard in the pursuit of person identification through chest X-ray images.

While a new paradigm of scientific research based on data centres and research infrastructures is gaining ground in science, and convergence between infrastructures and scientific domains is growing in cyberspace, epistemic cultures, particularly conservative in some fields, play a significant role in the dynamics of knowledge production in general and the adoption of data-intensive scientific practices in particular. In the present study, we focus on the transformations of scholarly communication through the perspective of digital curation of research data in the humanities, which certainly belong to these conservative epistemic cultures. The aim of this paper is to explore perspectives on the evolution of data curation in the context of the transformation of scholarly communication and research infrastructure in the humanities, specifically static archives, into living, continuously enriched data archives supported by artificial intelligence tools. To explore this perspective, we have chosen to compare scholarly communication in the humanities and in high-energy physics, in addition to analysing the practices of data curation itself. We further thematize the identified differences in terms of virtual research environments that can help humanities scholars exploit the potential of data-intensive research infrastructures.

The writing process, in which feelings and thoughts are expressed in writing, differs from person to person. Handwriting samples, which are very easy to obtain, are frequently used to identify individuals because they are biometric data. Today, with human-machine interaction increasing by the day, machine learning algorithms are frequently used in offline handwriting identification. Within the scope of this study, a dataset was created from 3250 handwritten images of 65 people. We tried to classify collected handwriting samples according to person and gender. In the classification made for person and gender recognition, feature extraction was done using 32 different transfer learning algorithms in the Python program. For person and gender estimation, the classification process was carried out using the random forest algorithm. 28 different classification algorithms were used, with DenseNet169 yielding the most successful results, and the data were classified in terms of person and gender. As a result, the highest success rates obtained in person and gender classification were 92.46% and 92.77%, respectively.

With the vast success of e-recruitment, online job offers have increased. Therefore, there is a number of job portals and recommendation systems trying to help users filter this massive amount of offers when searching for the right job. Until today, most of these systems' searching techniques are confined to using keywords such as job titles or skills, which also returns many results. This paper proposes a job recommender system that exploits the candidate's resume to select the appropriate job. Our system, AnnoJob, adopts a semantic annotation approach to: (1) intelligently extract contextual entities from resumes/offers, and (2) semantically structure the extracted entities in RDF triples using domain ontology, providing a unified presentation of the content of the documents. Furthermore, to select the suitable offer, we propose a novel semantic matching technique that computes the similarity between the resume/offers based on identifying the semantic similarity and relatedness between the RDF triples using the domain ontology and Wikidata, which enhance job-ranking results over existing information retrieval approaches. We evaluate our system using various experiments on data from real-world recruitment documents.

This paper deals with the prediction of company bankruptcies and defines how this undesirable state can be prevented. Currently, these methods include modern approaches from the area of data mining that can help companies in many ways. In a practical application of data mining methods for predicting the future state of a company, financial indicators of Polish companies were used. In the analyses, we used algorithms suitable for bankruptcy prediction – decision trees that provide a simple interpretation of results. In some experiments, we also used attribute selection methods, LASSO, or the PCA method. The workflow is governed by the CRISP-DM methodology, which describes the important steps needed for different analytical tasks. Part of the article is an analysis of the current state, which presents solutions to this problem suggested by other authors. After evaluating all models, we concluded that the C5.0 algorithm is capable of predicting a company’s bankruptcy or non-bankruptcy with 97.07 % accuracy, without the use of attribute selection methods.

Blackout events in smart grids can have significant impacts on individuals, communities and businesses, as they can disrupt the power supply and cause damage to the grid. In this paper, a new proactive approach to an early warning system for predicting blackout events in smart grids is presented. The system is based on deep learning models: convolutional neural networks (CNN) and deep self-organizing maps (DSOM), and is designed to analyse data from various sources, such as power demand, generation, transmission, distribution and weather forecasts. The system performance is evaluated using a dataset of time windows and labels, where the labels indicate whether a blackout event occurred within a given time window. It is found that the system is able to achieve an accuracy of 98.71% and a precision of 98.65% in predicting blackout events. The results suggest that the early warning system presented in this paper is a promising tool for improving the resilience and reliability of electrical grids and for mitigating the impacts of blackout events on communities and businesses.

Machine learning methods are used for purposes such as learning and estimating a feature or parameter sought from a dataset by training the dataset to solve a particular problem. The transfer learning approach, aimed at transferring the ability of people to continue learning from their past knowledge and experiences to computer systems, is the transfer of the learning obtained in the solution of a particular problem so that it can be used in solving a new problem. Transferring the learning obtained in transfer learning provides some advantages over traditional machine learning methods, and these advantages are effective in the preference of transfer learning. In this study, a total of 1980 eye contour images of 96 different people were collected in order to solve the problem of recognizing people from their eye images. These collected data were classified in terms of person, age and gender. In the classification made for eye recognition, feature extraction was performed with 32 different transfer learning algorithms in the Python program and classified using the RandomForest algorithm for person estimation. According to the results of the research, 30 different classification algorithms were used, with the ResNet50 algorithm being the most successful, and the data were also classified in terms of age and gender. Thus, the highest success rates of 83.52%, 96.41% and 77.56% were obtained in person, age and gender classification, respectively. The study shows that people can be identified only by eye images obtained from a smartphone without using any special equipment, and even the characteristics of people such as age and gender can be determined. In addition, it has been concluded that eye images can be used in a more efficient and practical biometric recognition system than iris recognition.

This paper presents a decision support system for identifying customer typology using cluster analysis to segment relevant customers. The approach is demonstrated using data from a company selling nutritional supplements, consisting of approximately 130,000 records from six Central European countries. The analysis results in distinct groups of customers, which are proposed for more effective management of customer relationships. The findings have implications for retailers, helping them focus on the most profitable customer segments to increase sales and profits and build lasting relationships. Furthermore, cluster analysis proves to be an appropriate statistical method for classification and provides valuable insights into patterns and trends in the analysed data. Overall, this paper contributes to development and comparison of methods for customer segmentation and demonstrates their potential for improving economic efficiency and building long-term customer relationships.

The pursuit of fairness in machine learning models has become increasingly crucial across various applications, including bank loan approval and face detection. Despite the widespread use of artificial intelligence algorithms, concerns persist regarding biases and discrimination within these models. This study introduces a novel approach, termed “The Fairness Stitch” (TFS), aimed at enhancing fairness in deep learning models by combining model stitching and training jointly, while incorporating fairness constraints. We evaluate the effectiveness of TFS through a comprehensive assessment using two established datasets, CelebA and UTKFace. The evaluation involves a systematic comparison with the existing baseline method, fair deep feature reweighting (FDR). Our analysis demonstrates that TFS achieves a better balance between fairness and performance compared to the baseline method (FDR). Specifically, our method shows significant improvements in mitigating biases while maintaining performance levels. These results underscore the promising potential of TFS in addressing bias-related challenges and promoting equitable outcomes in machine learning models. This research challenges conventional wisdom regarding the efficacy of the last layer in deep learning models for debiasing purposes. The findings suggest that integrating fairness constraints into our proposed framework (TFS) can lead to more effective mitigation of biases and contribute to fairer AI systems.

Stock forecasting is a complicated and daily challenge for investors because of the non-linearity of the market and the high volatility of financial assets such as stocks, bonds and other commodities. There is a need for a powerful and adaptive stock prediction model that handles complexities and provides accurate predictions. The support vector regression (SVR) model is one of the most prominent machine learning models for forecasting time series data. An ensemble hyperbolic tangent kernel SVR (HTK-SVR-BO) is proposed in this paper, combining Tanh and inverse Tanh kernels with Bayesian optimization. Combining the strengths of multiple kernels using the ensemble technique and then using optimization to identify the optimal values for each SVR model to enhance the ensemble model performance is possible. Our proposed model is compared with an ensemble SVR model (LPR-SVR-BO), which uses well-known SVR kernel types, including linear, polynomial and radial basis function (RBF). We apply the proposed models to Microsoft Corporation (MSFT) stock prices. The mean absolute error (MAE), mean squared error (MSE), root mean squared error (RMSE), R2 score (model accuracy) and mean absolute percentage error (MAPE) are the regression metrics used to compare the effectiveness of each ensemble model. In our comparison, HTK-SVR-BO performs better in terms of regression metrics compared to LPR-SVR-BO and achieves results of 0.27424, 0.13392, 0.36595, 0.99997 and 5.2331 respectively. According to the analysis, the proposed model is more predictive and may generalize to previously unknown data more effectively, so it can be accurate when forecasting future stock prices.

The increasing reliance on Low-power and Lossy Networks (LLN) in the Internet of Things (IoT) and their vulnerability to various attacks have made their protection necessary. Most of the proposed approaches to protecting such networks neither support scalability nor are lightweight enough to be incorporated into these constrained networks. In this paper, we present a lightweight approach to protecting LLN networks from rank and version number attacks. The proposal relies on light exchanged messages between the network nodes and the root and network nodes themselves. Successive comparison processes are used to detect the attack while two blocking techniques are introduced against malicious nodes. Simulations demonstrated the effectiveness of the approach, outperforming similar approaches such as Sink-Based Intrusion Detection Systems (SBIDS), Secure Routing Protocol (SRPL) for LLN, the Machine-Learning Technique based on K-Nearest Neighbour (MLTKNN), secure trust-aware Routing Protocol for Low Power and Lossy Networks (SecTrust) and Shield in terms of appropriate parameters, without causing extra charges on the network.

Blockchain is regarded as a significant innovation and shows a set of promising features that can certainly address existing issues in real time applications. Decentralization, greater transparency, improved traceability and secure architecture can revolutionize healthcare systems. With the help of advancement in computer technologies, most healthcare institutions try to store patient data digitally rather than on paper. Electronic health records are regarded as some of the most important assets in healthcare system and are required to be shared among different hospitals and other organizations to improve diagnosis efficiency. While sharing patients’ details, certain basic standards such as integrity and confidentiality of the information need to be considered. Blockchain technology provides the above standards with features of immutability and granting access to stored information only to authorized users. The examination approach depends on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (or PRISMA) rules and an efficient planned search convention is utilized to look through multiple scientific databases to recognize, investigate and separate every important publication. In this paper, we present a solid systematic review on the blockchain and healthcare domain to identify the existing challenges and benefits of applying blockchain technology in healthcare systems. More than 150 scientific papers published in the last ten years are surveyed, resulting in the identifications and summarization of observations made on the different privacy-preserving approaches and also assessment of their performances. We also present a significant architectural solutions of blockchain to achieve interoperability. Thereby, we attempt to analyse the ideas of blockchain in the medical domain, by assessing the advantages and limitations, subsequently giving guidance to other researchers in the area.

This editorial summarizes the special issue entitled “Theory and Practice of Generative Artificial Intelligence Usage” and briefly reflects on generative artificial intelligence-related phenomena. Generative artificial intelligence is a new technology that is not only known in professional circles but has also gained widespread recognition and popularity due to its easy accessibility and benefits for lay users. The high level of interest in this technology stems not only from current experience with it and its current usefulness, but also from the assumption that this new technology will significantly affect a wide range of industries, many types of jobs and the everyday lives of many people in the future. That is why, in this special issue, we have focused not only on the purely technical aspects of GenAI, but also on the social and commercial aspects related, for example, to institutional innovation, as well as ethical aspects and philosophical perspectives. This special issue consists of seven articles.

This study aims to analyse the role of bitcoin and gold as safe haven assets for Asian equity markets during periods of high market uncertainty related to the global COVID-19 pandemic, high volatility and extreme stock market conditions. Our empirical analysis employs the DCC-GARCH methodology to estimate the time-varying relationship between bitcoin/gold and the Asian stock market from 2016 to 2023. Our findings reveal that bitcoin serves as a strong hedge for Taiwan and Pakistan, whereas gold can be considered a strong hedge for Japan, Singapore, India, Thailand and Vietnam. Interestingly, we observe that bitcoin does not exhibit safe haven properties in any of the Asian countries observed. In contrast, gold demonstrates strong safe haven abilities for Singapore, India and Thailand. These results remain consistent across various measures of market turmoil, including the volatility index, COVID-19-related periods and low quantiles on the stock market. Furthermore, our results suggest that the perception and adoption of gold as a safe haven asset in Japan and Vietnam is mainly influenced by global events and uncertainties, rather than localized stock market conditions. These findings offer valuable information for investors, financial institutions as well as policy makers and regulators as to how cryptocurrency and gold have evolved as hedge and safe haven assets in Asia during uncertainty periods.

Deep learning has made identifying breast cancer lesions in mammography images an easy task in modern medicine, which has helped improve the diagnosis efficiency, sensitivity and accuracy by precisely identifying breast cancer from mammography images, contributing to timely detection and maintaining consistent performance. This paper presents the steps and strategies to develop a deep learning (DL) model to detect lesions in mammography images, based on U-Net architecture for precise segmentation, which has been developed for biomedical image segmentation, and incorporating ResNet34 as its encoder to extract features. Next, we employ the FastAI library, which simplifies and accelerates the model training tasks. For the data, studies and available resources lead us to INbreast, which is built with full-field digital mammograms contrary to other digitized mammograms. We obtained a high accuracy of 98% on the INbreast database, which is very challenging compared to state-of-the-art results.

In recent years, community detection is important because members of the same community share the same concepts. For efficient community detection in a social network, the influence node plays a vital role. A node in the social network or a user that has great influence and power would have a close relationship with a core of the group, termed a community. Therefore, the status of a person is determined by the user’s influence strength. That is, a user who has greater influence and strength plays a vital role in the social media community and also acts as a core in the community of the social network. Therefore, a community is a group of nodes in the complex network structure which are interlinked with one another. Effective community detection in a complex structure is a challenging task. Many studies have been done based on topological networks. The approaches are ineffective, inefficient and require more time to process. To overcome these issues, this paper proposes improving the influence nodes in complex networks by using the InfoMap with sigmoid fish swarm optimization algorithm (I-SFSO). Our proposed I-SFSO gives better accuracy rates for the data sets: 92% for Dolphin, 95% for the Facebook dataset, 96% for the Twitter data set, 94% for the YouTube data set, 93% for a karate club and 94% for football.

Osteosarcoma is a high-grade malignant bone tumour for which neoadjuvant chemotherapy is a vital component of the treatment plan. Chemotherapy brings about the death of tumour tissues, and the rate of their death is an essential factor in deciding on further treatment. The necrosis quantification is now done manually by visualizing tissue sections through the microscope. This is a crude method that can cause significant inter-observer bias. The suggested system is an AI-based therapeutic decision-making tool that can automatically calculate the quantity of such dead tissue present in a tissue specimen. We employ U-Net++ and DeepLabv3+, pre-trained deep learning algorithms for the segmentation purpose. ResNet50 and ResNet101 are used as encoder parts of U-Net++ and DeepLabv3+, respectively. Also, we synthesize a dataset of 555 patches from 37 images captured and manually annotated by experienced pathologists. Dice loss and Intersection over Union (IoU) are used as the performance metrics. The training and testing IoU of U-Net++ are 91.78% and 82.64%, and its loss is 4.4% and 17.77%, respectively. The IoU and loss of DeepLabv3+ are 91.09%, 81.50%, 4.77%, and 17.8%, respectively. The results show that both models perform almost similarly. With the help of this tool, necrosis segmentation can be done more accurately while requiring less work and time. The percentage of segmented regions can be used as the decision-making factor in the further treatment plans.

Model-Driven Engineering (MDE) is a software development paradigm that uses models as the main artifacts in the development process. MDE uses model transformations to propagate changes between source and target models. In some development scenarios, target models should be updated based on the evolution of source models. In such cases, it is required to re-execute all transformation rules to update the target model. Incremental execution of transformations, which partially executes the transformation rules, is a solution to this problem. The Epsilon Transformation Language (ETL) is a well-known model transformation language that does not support incremental executions. In this paper, we propose an approach to support the incremental execution of ETL transformations. Our proposal includes a process, as well as a prototype, to propagate changes to the target model. In the proposed approach, all the changes in the source model are detected to identify and re-execute the rules which deal with computing the required elements for updating the target model. We evaluated the correctness and performance of our approach by means of a case study. Compared to the standard ETL, the results are promising regarding the correctness of target models as well as faster execution of the transformation.

Classification is a crucial stage in identification systems, most specifically in biometric identification systems. A weak and inaccurate classification system may produce false identity, which in turn impacts negatively on delicate decisions. Decision making in biometric systems is done at the classification stage. Due to the importance of this stage, many classifiers have been developed and modified by researchers. However, most of the existing classifiers are limited in accuracy due to false representation of image features, improper training of classifier models for newly emerging data (over-fitting or under-fitting problem) and lack of an efficient mode of generating model parameters (scalability problem). The Negative Selection Algorithm (NSA) is one of the major algorithms of the Artificial Immune System, inspired by the operation of the mammalian immune system for solving classification problems. However, it is still prone to the inability to consider the whole self-space during the detectors/features generation process. Hence, this work developed an Optimized Negative Selection Algorithm (ONSA) for image classification in biometric systems. The ONSA is characterized by the ability to consider whole feature spaces (feature selection balance), having good training capability and low scalability problems. The performance of the ONSA was compared with that of the standard NSA (SNSA), and it was discovered that the ONSA has greater recognition accuracy by producing 98.33% accuracy compared with that of the SNSA which is 96.33%. The ONSA produced TP and TN values of 146% and 149%, respectively, while the SNSA produced 143% and 146% for TP and TN, respectively. Also, the ONSA generated a lower FN and FP rate of 4.00% and 1.00%, respectively, compared to the SNSA, which generated FN and FP values of 7.00% and 4.00%, respectively. Therefore, it was discovered in this work that global feature selection improves recognition accuracy in biometric systems. The developed biometric system can be adapted by any organization that requires an ultra-secure identification system.

This editorial summarises the special issue entitled “Future Trends of Machine Intelligence in Science and Industry”, which brings together several pieces of research that showcase the transformative impact of deep learning and intelligent systems across various domains, including healthcare, security and communication networks. By exploring advanced methodologies and innovative applications, this collection highlights significant strides in medical imaging, mental health diagnosis, biometric identification, smart grid management and adaptive e-learning. The featured articles delve into topics such as breast cancer detection using UNET architecture, psychodiagnosis prediction with deep learning, and blockchain-secured IoT systems for healthcare. Additionally, the issue covers revolutionary approaches in historical manuscript analysis, and contactless palm-print recognition. Through these comprehensive studies, we aim to inspire further advancements and cross-disciplinary collaborations, pushing the boundaries of what is achievable with modern technology.

The research herein applies text classification with which to categorize Khmer news articles. News articles were collected from three online websites through web scraping and grouped into nine categories. After text preprocessing, the dataset was split into training and testing sets. We then evaluated the performance of the ensemble learning method via machine learning classifiers with k-fold validation. Various machine learning classifiers were employed, namely logistic regression, Complement Naive Bayes, Bernoulli Naive Bayes, k-nearest neighbours, perceptron, support vector machines, stochastic gradient descent, AdaBoost, decision tree, and random forest were employed. Accuracy was improved for the categorization of Khmer news articles, in which Grid Search CV was used to find the optimal hyperparameters for each machine learning classifier with feature extraction TF-IDF and Delta TF-IDF. The results determined that the highest accuracy was achieved through the ensemble learning method in the support vector machine with the optimal hyperparameters (C = 10, kernel = rbf), using feature extraction TF-IDF and Delta TF-IDF, at 83.47% and 83.40%, respectively. The model establishes that Khmer news articles can be accurately categorized.

Stress is a part of everyone’s life. Any event or thought that makes you upset, furious or anxious can set it off. It will affect the human health mentally and physically and produce a negative impact on nervous and immune systems in our body. The human voice carries a lot of information about the person speaking. It also aids in determining a person's current state. In this proposed method, stress was detected using a deep learning model. Automatic stress detection is becoming an intriguing study topic as the necessity for communication between humans and intelligent systems rises. The hormone called cortisol can also be used to determine the body’s stress state. For most people, however, it is not a viable option. Speech features are particularly affected by stress, which is combined with the aim that voice data would serve as an easy-to-capture measure of everyday human stress levels and hence as an early warning signal of stress-related health problems. The proposed technique extracts Mel filter bank spectral coefficients from pre-processed voice input and the spectrum coefficients are extracted. The features of Mel frequency cepstral coefficients are applied to feed-forward networks and long short-term memory to predict the status of stress output using a binary decision, i.e., unstressed or stressed. The Mel spectrum and spectrogram output shows the variation in stressed and unstressed voice features. The results of the proposed method indicate better performance compared to an existing model. The model was developed as a web application to be used by workers to test their state of stress at any time.

Marketers utilize gamification as it provides an efficacious platform to communicate and reach a large consumer base. Previous literature has explored the impact of various aspects of gamification on consumer behaviour. This review synthesizes such studies and systematically examines 68 publications from Scopus from 2012 to 2022, employing bibliometric and systematic analysis. Performance analysis and science mapping are evaluated in bibliometric analysis to ascertain the most influential authors, documents, countries and journals. Thereafter, through cluster analysis, five major themes are identified, namely customer engagement, consumer experience, purchase and behavioural intentions, consumer psychology, and technology and innovation adoption. Additionally, in the systematic analysis, a conceptual framework is conceptualized depicting antecedents and consequences. Subsequently, practical implications and pertinent future directions are identified based on three aspects: context, method and theory.

Skin cancer is undoubtedly one of the deadliest diseases, and early detection of this disease can save lives. The usefulness and capabilities of deep learning in detecting and categorizing skin cancer based on images have been investigated in many studies. However, due to the variety of skin cancer tumour shapes and colours, deep learning algorithms misclassify whether a tumour is cancerous or benign. In this paper, we employed three different pre-trained state-of-the-art deep learning models: DenseNet121, VGG19 and an improved ResNet152, in classifying a skin image dataset. The dataset has a total of 3297 dermatoscopy images and two diagnostic categories: benign and malignant. The three models are supported by transfer learning and have been tested and evaluated based on the criteria of accuracy, loss, precision, recall, f1 score and ROC. Subsequently, the results show that the improved ResNet152 model significantly outperformed the other models and achieved an accuracy score of 92% and an ROC score of 91%. The DenseNet121 and VGG19 models achieve accuracy scores of 90% and 79% and ROC scores of 88% and 75%, respectively. Subsequently, a deep residual learning skin cancer recognition (ResNetScr) system has been implemented based on the ResNet152 model, and it has the capacity to help dermatologists in diagnosing skin cancer.

Medical document verification is a critical and expensive process that often relies on centralized databases. However, manual verification of such documents is time-consuming and lacks credibility. Deep learning and blockchain technology can be employed to address this issue by reducing fraud and increasing efficiency. The use of non-transferable soulbound tokens (SBTs) can provide a secure and tamper-proof system for verifying medical records. The authors have proposed an algorithm for automated document verification and authenticity using blockchain-based SBTs. The system uses cloud computing to access the decentralized database, reducing the time taken to verify each document to 2-3 minutes in comparison to the related non-automated techniques discussed in the literature review. The aim of this research paper is to provide a secure and tamper-proof system for verifying medical records, such as prescriptions and test results, on the cloud using decentralized databases and blockchain technology. The use of deep learning algorithms can be used to determine the best way to allocate resources in a decentralized network or to minimize the costs of a blockchain platform. The adoption of blockchain technology can reduce fraud and improve efficiency. The proposed system can significantly improve the efficiency and credibility of medical document verification, reduce fraud, and ensure tamper-proof authenticity. The use of SBTs and cloud computing can simplify the process and provide easy access to decentralized databases. Future research can explore the scalability of the proposed system and its potential application in other sectors.

Social media, blogs, review sites and forums can produce large volumes of data in the form of users’ emotions, views, arguments and opinions about various political events, brands, products and social problems. The user's sentiment expressed on the web influences readers, politicians and product vendors. These unstructured social media data are analysed to form structured data, and for this reason sentiment analysis has recently received the most important research attention. Sentiment analysis is a process of classifying the user’s feelings in different manners such as positive, negative or both. The major issue of sentiment analysis is insufficient data processing and outcome prediction. For this, deep learning-based approaches are effective due to their autonomous learning ability. Emotion identification from the text in natural language processing (NLP) provides more benefits in the field of e-commerce and business environments. In this paper, emotion detection-based text classification is used for sentiment analysis. The data collected are pre-processed using tokenization, stop word discarding, stemming and lemmatization. After performing data pre-processing, the features are identified using term frequency and inverse document frequency (TF-IDF). Then the filtered features are turned into word embeddings by documents as a vector (Doc2Vec). Then, for text classification, a deep learning (DL) based model called convolutional bidirectional long short-term memory (CBLSTM) is used to differentiate the sentiments of human expression into positive or good and negative or bad emotions. The neural network hyper-parameters are optimized with a meta-heuristic algorithm called the frog leap approach (FLA). The proposed CBLSTM with FLA uses four review and Twitter datasets. The experimental results of this study are compared with the conventional approaches LSTM-RNN and LSTM-CNN to prove the efficiency of the proposed model. Compared to LSTM-RNN and LSTM-CNN, the proposed model secures an improved average accuracy of 98.1% for review datasets and 97.5% for Twitter datasets.

This paper introduces the praxeologically grounded research programme on socio-informatics developed at the University of Siegen and the International Institute for Socio-Informatics in Germany. We outline our methodological framework of grounded design, guiding and binding together a variety of different engagements in practice which we call design case studies (DCS). While grounded design is applied to a broad variety of different areas, in this paper we deal specifically with our engagements in support of communities and political activists. To exemplify our approach, we present one DCS: It focuses on the participatory development of a computer club in the High Atlas as a facilitation infrastructure to support public community participation in the process of modernization and social transformation in the Maghreb. Insights into the project context, the objectives and methodology and the depiction of some of its first findings lays the ground for the discussion of socio-informatics research principles.

The features of human speech signals and emotional states are used to estimate the blood pressure (BP) using a clustering-based model. The audio-emotion-dependent discriminative features are identified to distinguish individuals based on their speech to form emotional groups. We propose a bio-inspired Enhanced grey wolf spotted hyena optimization (EWHO) technique for emotion clustering, which adds significance to this research. The model derives the most informative and judicial features from the audio signal, along with the person’s emotional states to estimate the BP using the multi-class support vector machine (SVM) classifier. The EWHO-based clustering method gives better accuracy (95.59%), precision (97.08%), recall (95.16%) and F1 measure (96.20%), as compared to other methods used for BP estimation. Additionally, the proposed EWHO algorithm gives superior results in terms of parameters such as the silhouette score, Davies-Bouldin score, homogeneity score, completeness score, Dunn index, and Jaccard similarity score.