Fulltext search in archive

In recent years, cloud computing has been widely used in various fields and is gaining importance in healthcare systems. Patients’ health data are outsourced to cloud storage, enabling healthcare professionals to easily access health information from anywhere and at any time to improve health services. Once patient data are stored in the cloud, they are vulnerable to attacks such as data loss, denial of service (DoS), distributed denial of service (DDoS) and other sorts of cyberattacks. Data confidentiality and patient privacy are more of a problem in the cloud computing context due to their public availability. If a patient's personal information is stolen, he or she may face a range of problems. These are concerns that necessitate more security. The transmission of this sensitive information over the internet is always susceptible to hacking. Therefore, the privacy of patients’ data is considered one of healthcare organizations’ main issues. To overcome this problem, encryption mechanisms that place a significant emphasis on securing data within the cloud environment are used to preserve sensitive health data. A hybrid cryptography approach is employed in this paper to ensure the secure sharing of health data over the cloud. To maintain data privacy and secrecy, a hybrid cryptography mechanism for storing and transporting data to and from the cloud is used. To protect data from malevolent insiders, the encryption key is separated into two halves, controlling access to patient records via a specific technique. This paper shows the implementation and performance evaluation of the proposal as a functional system prototype. The evaluation is based on the key generation time, the record encryption time, the record decryption time, the record upload time and the record download time for different user numbers and different file sizes varying from 0.1 MB to 500 MB. The findings show that the proposal performs better than other state-of-the-art systems and can practically share secure health data in cloud environments.

Sharing of Electronic Health Records (EHRs) is of significant importance in health care. Lately, a cloud-based electronic health record sharing scheme has been used extensively to share patient records among various healthcare organizations. However, cloud centralization may compromise patients’ privacy and security. Due to the special features of blockchain, it is important to see this technology as a promising solution to resolve these issues. This article proposes a privacy-preserving, secure EHR sharing and access control framework based on blockchain technology. The proposal aims to implement EHR blockchain technology and ensure that electronic records are stored safely by specifying user access permissions. We emulate the cryptographic primitives and use smart contracts to describe the relationships between the EHR owner and EHR user through the proposed system on the Ethereum blockchain. We assess the proposal results based on encryption and decryption time and the costs of the smart contract. The encryption and decryption times are proportional to the size of the EHR, which varies from 128 KB to 128 MB. When it comes to encryption, the smallest EHR takes 0.0012 s to encrypt, while the largest EHR, which is 128 MB, takes 1.4149 s. On the other hand, a 128 KB EHR takes 0.0013 s to decrypt, whereas a 128 MB EHR requires 1.6284 s. As a result, performance evaluation and security analysis confirm that the proposal is secure for practical application.

Alzheimer's disease is a brain memory loss disease. Usually, it will affect persons over 60 years of age. The literature has revealed that it is quite difficult to diagnose the disease, so researchers are trying to predict the disease in the early stage. This paper proposes a framework to classify Alzheimer's patients and to predict the best classification algorithm. The Bestfirst and CfssubsetEval methods are used for feature selection. A multi-class classification is done using machine learning algorithms, namely the naïve Bayes algorithm, the logistic algorithm, the SMO/SMV algorithm and the random forest algorithm. The classification accuracy of the algorithms is 67.68%, 84.58%, 87.42%, and 88.90% respectively. The validation applied is 10-fold cross-validation. Then, a confusion matrix is generated and class-wise performance is analysed to find the best algorithm. The ADNI database is used for the implementation process. To compare the performance of the proposed model, the OASIS dataset is applied to the model with the same algorithms and the accuracy of the algorithms is 98%, 99%, 99% and 100% respectively. Also, the time for the model construction is compared for both datasets. The proposed work is compared with existing studies to check the efficiency of the proposed model.

This paper presents two novel deterministic initialization procedures for k-means clustering based on a modified crowding distance. The procedures, named ck-means and fck-means, use more crowded points as initial centroids. Experimental studies on multiple datasets demonstrate that the proposed approach outperforms k-means and k-means++ in terms of clustering accuracy. The effectiveness of ck-means and fck-means is attributed to their ability to select better initial centroids based on the modified crowding distance. Overall, the proposed approach provides a promising alternative for improving k-means clustering.

Studying the ways in which people construct their identities in online environments is a pressing contemporary concern. The research reported in this article was designed to examine the uses of, and influences on, Vietnamese respondents’ identity formation on Facebook. Data were collected by means of a social survey and the application of the Zaltman metaphor elicitation technique, a procedure that searches for customers’ thoughts and emotions by digging deep into the visual and non-visual illustrations that customers collect or make on their own. The findings show how Vietnamese Facebook users present themselves and how they thereby facilitate their self-expansion and maintain their sense of self-esteem. According to the analysis, it can be suggested that Facebook is where adults portray their socially conformed versions against social reflection. They use this platform to seek validation, demonstrate their professional side to make them look better in the eyes of society. Drivers of online identity formation are revealed through negotiating with conflicts in their existing identities. The level of self-modification amongst respondents is adjusted in relation to their social vigilance, conformity and motivations in and between social categorization.

The article presents a focused analysis of the Russian-language scientific literature of the first decades of the twenty-first century on the problems of the development of social informatics in Russia. The authors have shown that the multidimensionality of social informatics causes an increasing interdisciplinary research interest in the professional community and is divided into four directions. First, researchers refined the conceptual foundations of social informatics. The second direction is devoted to the study of information resources in their dynamics. The third direction considers and analyses digital traces and their use under the conditions of digital transformation. The fourth direction is devoted to the problems of children and young people in the information environment. The authors achieve intermediate results. Firstly, the clarification of the conceptual basis of the study of social informatics is due to the public demand for comprehensive informatization. Secondly, interest in information resources is due to the need for effective organization of data series in information systems. Thirdly, the accounting and analysis of digital traces and their use are due to the public need for collective security. Fourthly, it is necessary to observe the problems of perception and use of digital technologies by children and young people, as well as their socialization in the developing information society. Finally, the authors conclude that the potential of social informatics is increased by an interdisciplinary interest in qualitative changes in Russian society in the context of digital transformation.

The proliferation of smartphones and sensor-based networks has led to a greater need for context-aware applications and pervasive business processes. One of the key approaches that seek to satisfy this need is context-aware service composition. Service composition can be achieved in two ways, i.e., service choreography and service orchestration. Embedding the context into an orchestrated composite service enhances its flexibility, but makes its development and execution more complicated. This study aims to reduce this complexity by introducing the CA-BPEL approach. Our proposed approach enables developers to turn a standard orchestrated service into a context-aware orchestrated service, consistent with the standard WS-BPEL language. This study applies the Design Science Research Methodology, in which we evaluate CA-BPEL by using a tourism demonstration along with the conduction of a usability survey that shows the convenience of the proposed approach. We also compare our proposed approach with 14 related studies. Our investigations suggest that CA-BPEL has much potential to facilitate the development and execution of context-aware service compositions.

This paper analyses data from the real estate market domain. The data were scraped from the bezrealitky.cz portal. The analysis looks at both sales and rental data. A total of 3546 records and 54 attributes were obtained. A basic overview of the data was performed using exploratory data analysis where some basic characteristics of the data were identified, such as the average price of sold and rented flats. More specific results were obtained by applying data mining methods such as regression (linear regression, lasso regression and ridge regression) for predicting the flat prices and payments for utilities, classification (support vector machines, KNN, Gaussian naïve Bayes, decision tree and random forest) for estimating the PENB class (building energy performance certificate) and building condition. Lasso regression performed the most successfully (R² = 0.76) in predicting the rent price. Among the classification tasks, the best result was achieved with random forest, which had an accuracy over 80% in some cases. Other tasks included clustering (k-means and k-modes) and anomaly detection (isolation forest). The main focus was on descriptive data mining, especially on clustering. Clusters created using the k-means algorithm (silhouette score of 0.78) with flats based on geographic coordinates were identified which show that the most expensive flats are on average in Bohemian regions, followed by Silesia and the cheapest are in central Moravia. Another cluster application identified flats in the Moravian-Silesian region with very high payments for utilities (silhouette score of 0.56). The models can help estimate the value of flats based on their attributes as well as location.

The COVID-19 coronavirus illness is caused by a newly discovered species of coronavirus known as SARS-CoV-2. Since COVID-19 has now expanded across many nations, the World Health Organization (WHO) has designated it a pandemic. Reverse transcription-polymerase chain reaction (RT-PCR) is often used to screen samples of patients showing signs of COVID-19; however, this method is more expensive and takes at least 24 hours to get a positive or negative response. Thus, an immediate and precise method of diagnosis is needed. In this paper, chest X-rays will be utilized through a deep neural network (DNN), based on a convolutional neural network (CNN), to detect COVID-19 infection. Based on their X-rays, those with COVID-19 indications may be categorized as clean, infected with COVID-19 or suffering from pneumonia, according to the suggested CNN network. Sample pieces from every group are used in experiments, and categorization is performed by a CNN. While experimenting, the CNN-derived features were able to generate the maximum training accuracy of 94.82% and validation accuracy of 94.87%. The F1-scores were 97%, 90% and 96%, in clearly categorizing patients afflicted by COVID-19, normal and having pneumonia, respectively. Meanwhile, the recalls are 95%, 91% and 96% for COVID-19, normal and pneumonia, respectively.

To be able to play an active role in the design, creation and development of a networked society, students, scholars and practitioners need basic knowledge in social informatics. Students at a university of applied sciences attended a one-term seminar that consisted of eight two-day workshops. The students learned and used theories, concepts and methods of social informatics (SI), focusing on the sociological part of SI. The learning and teaching approach is grounded in competency-based learning. It enables students to explore a certain field. This is implemented by means of organization development, project-based learning, agile learning and teaching plus blended learning. It empowers teachers and students to work together efficiently, effectively and with joy. To learn how and why this approach worked, an embedded case study investigated the design, implementation and evaluation of the workshop series. Finally, the paper describes impediments to the communication and understanding of the term and the field of social informatics and sketches ideas how to deal with it concerning study programmes and the value social informatics can contribute to challenges like a pandemic.

Automatic gender recognition is one of the frequently solved tasks in computer vision. It is useful for analysing human behaviour, intelligent monitoring or security. In this article, gender is recognized based on multispectral images of the hand. Hand (palm and back) images are obtained in the visible spectrum and thermal spectrum; then a fusion of images is performed. Some studies say that it is possible to distinguish male and female hands by some geometric features of the hand. The aim of this article is to determine whether it is possible to recognize gender by the thermal characteristics of the hand and, at the same time, to find the best architecture for this recognition. The article compares several algorithms that can be used to solve this issue. The convolutional neural network (CNN) AlexNet is used for feature extraction. The support vector machine, linear discriminant, naive Bayes classifier and neural networks were used for subsequent classification. Only CNNs were used for both extraction and subsequent classification. All of these methods lead to high accuracy of gender recognition. However, the most accurate are the convolutional neural networks VGG-16 and VGG-19. The accuracy of gender recognition (test data) is 94.9% for the palm and 89.9% for the back. Experiments in comparative studies have had promising results and shown that multispectral hand images (thermal and visible) can be useful in gender recognition.

This editorial summarises the special issue entitled “Deep Learning Blockchain-enabled Technology for Improved Healthcare Industrial Systems”, which deals with the intersection and use of deep learning and blockchain technologies in the healthcare industry. This special issue consists of eleven scientific articles.

The article deals with the history of the formation, current state and prospects for the development of social informatics as a current direction in science and education in Russia. The article offers mainly a personal view of the author, who has been involved in shaping social informatics in Russia for the last three decades. The article presents the distinctive features of the Russian scientific school of social informatics and its priorities in the formation of this field. The main directions of research in the field of social informatics in Russia in the context of the formation of the global information society are determined. Emphasis is placed on the interdisciplinary nature of many issues related to social informatics and their systematic study. Finally, the article summarizes the current necessity for the deep study of issues related to social informatics, e.g., information inequality, information crime, cyberbullies, manipulation of consciousness, virtualization of society, information wars, information poverty, information culture, and using computers to analyse social phenomena such as communication via social media. It is important not only in the area of scientific research but also in the system of secondary and higher education and training of scholars.

Low qualification of employees newly hired to service desks contributes to the high turnover of service desk agents and consequently to low quality of services delivered. This paper proposes a conceptual artefact comprising two modules for tacit knowledge elicitation and knowledge transfer during the onboarding training process. The design of the artefact follows the design science methodology. Ex-ante evaluation methods are chosen to evaluate the importance of a problem domain and evaluate the artefact feasibility. Expert interviews and focus group discussions with experts from the field were performed to support the evaluation activities. The proposed framework uses eye-tracking technology to complement captured knowledge with tacit knowledge. Next, the proposed model incorporates a simulated environment for enhanced training experience and effective knowledge transfer from expert employees to novice ones. This paper and the proposed artefact aim to improve the training process of service desk employees and to contribute to wider use of tacit knowledge capture and elicitation techniques in IT service management.

The increase in data diversity and the fact that database design is a difficult process make it practically impossible to design a unique database schema for all datasets encountered. In this paper, we introduce a fully automatic genetic algorithm-based relational database normalization method for revealing the right database schema using a raw dataset and without the need for any prior knowledge. For measuring the performance of the algorithm, we perform a simulation study using 250 datasets produced using 50 well-known databases. A total of 2500 simulations are carried out, ten times for each of five denormalized variations of all database designs containing different synthetic contents. The results of the simulation study show that the proposed algorithm discovers exactly 72% of the unknown database schemas. The performance can be improved by fine-tuning the optimization parameters. The results of the simulation study also show that the devised algorithm can be used in many datasets to reveal structs of databases when only a raw dataset is available at hand.

When interacting with each other virtually on social media sites, users may potentially encounter their own “self-presence”. To improve their physical self-presence on social media, users might resolve to curate themselves to match their ideal perceptions of themselves and others. This study examines the intention to alter one’s factual self in real life whilst contemplating the on-line self on social media. Data were derived from a social survey of Vietnamese adults. The results indicate the remarkable mediating relationship from the level of editing self-images to three choices of intention to change the body image in real life. Nevertheless, this paper does not provide enough evidence to confirm any link connecting distal intentions (D-intentions), proximal intentions (P-intentions) and motor intentions (M-intentions) with private self-consciousness, with the latter as the moderator variable. This study, therefore, might provide an exclusive idea of how online behaviour can be related to one’s offline behaviour in terms of body image. With an understanding of this aspect, many implications can be found in regard to applications/programmes in research, development and marketing.

A model is introduced to investigate the effect of social media application qualities on the using of these applications for daily news. A standard questionnaire was designed and distributed among randomly selected social media users in the city of Neyshabur in Iran. The content of the questionnaire was validated by experts and its reliability verified using Cronbach's alpha. Random sampling was used identify participants. The SmartPLS software was used to investigate the research findings. Structural equation modelling is used for data analysis. The results show that the factors of system quality, information quality, service quality and personalization of applications affect their perceived usefulness. System quality and service quality affect the perceived ease of use of applications. However, information quality does not affect perceived ease of use. Also, the results show that perceived usefulness has a greater effect (with a path coefficient of 45%) on the attitude in respect of perceived ease of use (path coefficient of 26%). Personalization has the most positive impact on perceived usefulness and service quality has a great impact on perceived ease of use. Facilitating conditions have a positive impact on the use of social media and their use for news. Furthermore, the results show that the factors that affect the use of applications in Telegram are better than those in Viber. These findings justify why Iranian people migrate from Viber to Telegram as their social media application.

The paper deals with an analysis and evaluation of an innovative visual design of the Czech Financial Administration application, which was launched in order to solve needs arising from the COVID-19 pandemic. The Financial Administration had not made use of information technologies for several years, which had prevented the streamlining of public services offered to taxpayers. The aim of the study is to compare the perception of the visual interface design of two Financial Administration applications: (a) the Tax Portal – launched before the pandemic, and (b) the application for the provision of a compensation bonus for self-employed – launched during the pandemic. The research is based on a sequential mixed method design, where findings from a focus group and an interview are used to define relevant properties of web design to be evaluated in a questionnaire survey. The difference between variables regarding the perception of the Financial Administration applications is determined using a paired t-test. The results show that the new application has brought a significant change in the appearance of the interface design. Positive results of the Financial Administration’s innovative approach can be beneficial for future development of different e-government projects.

With the COVID-19 pandemic experiences of Filipino students, the face-to-face mode of instruction in the classroom has been phased out in exchange for online learning platforms such as Google Classroom (GCR) among some K-12 learners. As advised by the Commission on Higher Education (CHED), most colleges and universities had to try available learning management systems; hence, this research study aimed to investigate the role of trialability and purposefulness towards GCR acceptance among tertiary institutions following the CHED educational policy. The researchers came up with eight hypotheses, which suggested that purposefulness may influence educational policy and acceptance of GCR. Trialability of GCR may influence educational policy and technical access. One thousand sixty-six (1066) respondents from six public higher institutions of learning were given online questionnaire; however, only 913 users were considered for the structural equation modelling and indirect effect of the suggested factors in this study. Using SmartPLS 3.0, the findings revealed that except for the hypothesis on institutional willingness (p< 0.054), all the hypotheses were highly supported at the level of significance p < 0.00 to p< 0.005. Thus, this study proves that GCR is an appropriate platform for colleges and universities. Trialability and purposefulness are two great factors that contributed to the acceptance and adoption of GCR in higher institutions of learning. Future researchers are therefore encouraged to replicate this study and validate the findings since the use of GCR is relatively new among Filipino teachers and learners.

In the current COVID-19 pandemic, teachers and students routinely take part in schooling via digital technologies, using their own and other authors' copyrighted works, even without their permission or licence. The aim of this article is to identify these aspects of copyright in online teaching and to offer solutions to them. Some of these aspects are addressed by EU Directive No. 2019/790 (Digital Directive), the implementation of which was reflected in the draft amendment to the Czech Copyright Act, which had not addressed them. To fulfil this goal, the author asked several research questions: What forms of use of copyright works refer to the terms e-learning, online teaching or digital teaching? Is it possible to apply a statutory teaching exemption to the use of copyright works of other authors without their consent under the current wording of the Copyright Act (AutZ)? Is there any case law regarding such an exemption that would be applicable to e-learning? How will the conditions for the use of copyright works in e-learning change with the AutZ amendment following the implementation of EU Directive No. 2019/790 (Digital Directive) and what practical organizational recommendations can be made? To the above goal and questions, we conducted research in Czech law relating to the use of copyright works in digital teaching. For this purpose, we used the method of empirical research of Czech legislation and case law containing the terms e-learning and online or digital teaching. We also employed a comparative method to compare Czech law and EU law. The article is primarily an overview with elements of reflection. It does not comprehensively address all aspects of copyright in teaching, such as plagiarism in general, but identifies copyright aspects specific to e-learning.

This article analyzes ways of calculating characteristics of data and most common data structure types that allow comparison between them or on a time axis. To achieve this, it studies the key aspects of relational databases, XML, JSON and RDF structure types. These data structure types are compared to multiple isolated approaches to data quality and other data characteristics measurements. The goals of the article are the calculation method itself and a storage structure for calculated values. The article presents a method of characterization of data and data structure types based on the calculation of three metrics: the amount of structuredness, the amount of hierarchicallity and the amount of information. This triad of metrics allows comparison between various data sets (objects), for example evaluating the complexity of the transformation of data from one data object to another, as well as with data structure types (as mentioned above). Based on the vector of three metrics, the calculation method of the compatibility between data and data structure type is proposed. This method can help select the most compatible data format for existing data. The calculated values of metrics can also detect non-optimal storage design and classify data transformations. The method was evaluated on an example case study, which showed its usability on an example demonstration data set. It can be used in the process of data modelling to help select optimal data structure type, to design a data transformation process and to optimize existing data storages.

Information and communication technologies are increasingly interacting with modern societies. One specific manifestation of this interaction concerns hateful and other negative comments in online environments. Various terms appear to denote this communication, from flaming, indecency and intolerance to hate speech. However, there is still a lack of an umbrella term that broadly captures this communication. Therefore, this paper introduces the concept of socially unacceptable discourse, which serves as the basis for an empirical study that evaluated online comments scraped from the Facebook pages of the three most-visited Slovenian news outlets. Machine-learning algorithms were used to narrow the focus to topics related to refugees and LGBT rights. Ten thousand comments were manually coded to identify and structure socially undesirable discourse. The results show that about half of all comments belonged to this type of discourse, with a surprisingly stable level and structure across media (i.e., right-wing versus mainstream) and topics. Most of these comments could also be considered a potential violation of hate speech legislation. In the context of these findings, the political and ideological consequences and implications of mediatised emotions are discussed.

Companies that use robotic process automation very often deal with the problem of selecting a suitable process for automation. Manual selection of a suitable process is very time-consuming. Therefore, part of the process mining field specializes in selecting suitable processes for automation based on process data. This work deals with the possibility of improving the existing method for finding suitable candidates for automation. To improve the current approach, we remove the limiting restrictions of the current method and use another FURIA rule-learning algorithm for rule detection. We use three different datasets and the WEKA platform to validate the results. The results show that FURIA and the removal of strictly deterministic rules as restrictions turned out to be a competitive approach to the original one. On data presented in this study, the selected approach detected more candidates for automation and with higher accuracy. This study implies that FURIA and not using a strictly deterministic process is an appropriate procedure with certain use cases as other procedures mentioned in this study.

The third most common cause of death worldwide is attributed to pulmonary diseases, making it imperative to diagnose them promptly. Radiology is a medical discipline that utilizes medical imaging to guide treatment. Radiologists prepare reports interpreting details and findings analysed from medical images. Radiology free-text reports are a rich source of textual information that can be exploited to enhance the efficacy of medical prognosis, treatment and research. Radiology reports exist in an unstructured format as are not suitable by themselves for mathematical computation or machine learning operations. Therefore, natural language processing (NLP) strategies are employed to convert unstructured natural language text into a structured format that can be fed into machine learning (ML) or deep learning (DL) models for information extraction. We propose a DL-based medical text modelling framework incorporating a knowledge base to predict pulmonary diseases from unstructured radiology free-text reports. We make detailed diagnostic performance evaluations of our proposed technique by comparing it with state-of-the-art NLP techniques on radiology free-text reports extracted from two medical institutions. The comprehensive analysis shows that the proposed model achieves superior results compared to existing state-of-the-art text modelling techniques.

The mortality due to stroke is increasing. Accurate prediction of stroke-caused death is very important for healthcare. Data mining methods are novel ways to predict these mortality risks. The aim of this study is to employ popular data mining algorithms to predict the survival of stroke patients and extract decision rules. The data on stroke patients (n=4149) were collected from paper medical records. Missing data were managed using the multiple imputation method. Also, the target variable was balanced using methods such as over-sampling, under-sampling and Synthetic Minority Oversampling (SMOTE). The support vector machine (SVM), decision tree, and logistic regression (LR) algorithms were employed to predict the survival of stroke patients. Also, the Repeated Incremental Pruning to Produce Error Reduction (RIPPER) algorithm was used to extract the decision rules from the main dataset. LR outperformed other algorithms in terms of accuracy (76.96%), sensitivity (79.06%) and kappa (33.34). However, specificity (65.35%) and AUC (0.77) scores were lower than those of other algorithms. An independent dataset with 234 records was selected to challenge the LR algorithm with the best performance from the main dataset. After employing this algorithm on the external validation dataset, its performance was improved in accuracy (79.91%), sensitivity (83.94%), kappa (39.26) and AUC (0.8), but not in specificity (60.98%). The constructed model predicted the survival of stroke patients with high scores and useful rules were extracted for clinical usage.

Distributed denial of service (DDoS) attacks are one of the most common global challenges faced by service providers on the web. It leads to network disturbances, interruption of communication and significant damage to services. Researchers seek to develop intelligent algorithms to detect and prevent DDoS attacks. The present study proposes an efficient DDoS attack detection model. This model relies mainly on dimensionality reduction and machine learning algorithms. The principal component analysis (PCA) and the linear discriminant analysis (LDA) techniques perform the dimensionality reduction in individual and hybrid modes to process and improve the data. Subsequently, DDoS attack detection is performed based on random forest (RF) and decision tree (DT) algorithms. The model is implemented and tested on the CICDDoS2019 dataset using different data dimensionality reduction test scenarios. The results show that using dimensionality reduction techniques along with the ML algorithms with a dataset containing high-dimensional data significantly improves the classification results. The best accuracy result of 99.97% is obtained when the model operates in a hybrid mode based on a combination of PCA, LDA and RF algorithms, and the data reduction parameter equals 40.

COVID-19 causes a jarring impact on the livelihoods of people in Malaysia and globally. To prevent an outbreak in the community, identifying the likely sources of infection (hotspots) of COVID-19 is important. The goal of this study is to formulate a bipartite network model of COVID-19 transmissions by incorporating patient mobility data to address the assumption on population homogeneity made in the conventional models and focus on indirect transmission. Two types of nodes – human and location – are the main concern in the research scenario. 21 location nodes and 31 human nodes are identified from a patient’s pre-processed mobility data. The parameters used in this study for location node and human node quantifications are the ventilation rate of a location and the environmental properties of the location that affect the stability of the virus such as temperature and relative humidity. The summation rule is applied to quantify all nodes in the network and the link weight between the human node and the location node. The ranking of location and human nodes in this network is computed using a web search algorithm. This model is considered verified as the error obtained from the comparison made between the benchmark model and the COVID-19 bipartite network model is small. As a result, the higher ranking of the location is denoted as a hotspot in this study, and for a human node attached to this node will be ranked higher in the human node ranking. Consequently, the hotspot has a higher risk of transmission compared to other locations. These findings are proposed to provide a framework for public health authorities to identify the sources of infection and high-risk groups of people in the COVID-19 cases to control the transmission at the initial stage.

Increasing demand for personalized tours for tourists travel in an urban area motivates more attention to points of interest (POI) and tour recommendation services. Recently, the granularity of POI has been discussed to provide more detailed information for tour planning, which supports both inside and outside routes that would improve tourists' travel experience. Such tour recommendation systems require a predefined POI database with different granularities, but existing POI discovery methods do not consider the granularity of POI well and treat all POIs as the same scale. On the other hand, the parameters also need to be tuned for different cities, which is not a trivial process. To this end, we propose a city adaptive clustering framework for discovering POIs with different granularities in this article. Our proposed method takes advantage of two clustering algorithms and is adaptive to different cities due to automatic identification of suitable parameters for different datasets. Experiments on two real-world social image datasets reveal the effectiveness of our proposed framework. Finally, the discovered POIs with two levels of granularity are successfully applied on inner and outside tour planning.

Recommender systems (RS) are designed to eliminate the information overload problem in today's e-commerce platforms and other data-centric online services. They help users explore and exploit the system's information environment utilizing implicit and explicit data from internal e-commerce systems and user interactions. Today's product catalogues include pictures to provide visual detail at a glance. This approach can effectively convert potential buyers into customers. Since most e-commerce stores use product images to promote, arouse users' visual desires and encourage them to buy products, this paper develops an image-based RS using deep learning techniques. To perform the research, we use five convolutional neural network (CNN) models to extract the features of the products' images. Then, the system uses the features to calculate the similarity between images. The selected CNN models are VGG16, VGG19, ResNet50, Inception V3 and Xception. We also analysed four versions of the MovieLens dataset to demonstrate the accuracy improvement of the recommendations, including 100k, 1M, 10M and 20M. Results of the experiment showed a significant increase in accuracy compared with traditional approaches. Also, we express many related open issues including use of multiple images per item, different similarity metrics, other CNN models, and the hybridization of image-based and different RS techniques for future studies. This method also provides more accurate product recommendations on e-commerce platforms than traditional methods.

Skin cancer is a significant healthcare problem with a high mortality rate worldwide. Skin lesions occur due to the abnormal growth of skin cells in humans. Failure of early prediction and proper lesion diagnosis may lead to a malignant stage. In recent times, different skin lesion images have appeared with high similarity. Hence, classification is a more challenging task with imbalances in the dataset. The proposed work is implemented as a hybrid model with a dynamic salp swarm algorithm (DSSA) with a weighted extreme learning machine (DSSA-WELM) that addresses the imbalances in the dataset and performs the classification with higher accuracy. GoogleNet is a pre-trained network model used with the hybrid model, which helps converge faster with the optimization process. The extreme learning machine (ELM) is a multiclass classifier for accurate dermoscopic image classification. The DSSA, the best feature selection algorithm enhances the classification accuracy of the WELM. Image classification is accomplished with the International Skin Imaging Collaboration 2019 benchmark dataset. The proposed solution classifies images into eight classes: melanoma, melanocytic nevus, basal cell carcinoma, actinic keratosis, dermatofibroma, vascular lesion, squamous cell carcinoma and unknown lesion. The efficiency of the proposed solution is proved by comparing it with various state-of-the-art approaches such as support vector machine (SVM), ELM, and particle swarm optimization (PSO) methods. Results are evaluated using standard metrics of sensitivity, specificity and precision. The proposed solution outperforms all these older approaches.