Search

Background: Cancer is a deadly disease. It is crucial to diagnose cancer in its early stages. This can be done with medical imaging. Medical imaging helps us scan and view internal organs. The analysis of these images is a very important task in the identification and classification of cancer. Over the past years, the occurrence of cancer has been increasing, so has been the load on the medical fraternity. Fortunately, with the growth of Artificial Intelligence in the past decade, many tools and techniques have emerged which may help doctors in the analysis of medical images. Methodology: This is a systematic study covering various tools and techniques used for medical image analysis in the field of cancer detection. It focuses on machine learning and deep learning technologies, their performances, and their shortcomings. Also, the various types of imaging techniques and the different datasets used have been discussed extensively. This work also discusses the various preprocessing techniques that have been performed on medical images for better classification. Results: A total of 270 studies from 5 different publications and 5 different conferences have been included and compared on the above-cited parameters. Conclusion: Recommendations for future work have been given towards the end.

The deep web is comprised of a large corpus of information hidden behind the searchable web interfaces. Accessing content through searchable interfaces is somehow a challenging task. One of the challenges in accessing the deep web is automatically filling the searchable web forms for retrieving the maximum number of records by a minimum number of submissions. The paper proposes a methodology to improve the existing method of getting informative data behind searchable forms by automatically submitting web forms. The form text field values are obtained through Bayesian inferences. Using Bayesian networks, the authors aim to infer the values of text fields using the existing values in the label value set (LVS) table. Various experiments have been conducted to measure the accuracy and computation time taken by the proposed value selection method. It proves to be highly accurate and takes less computation time than the existing term frequency-inverse document frequency (TF-IDF) method, hence increasing the performance of the crawler.

This paper proposes a novel algorithm based on reinforcement learning-entitled asynchronous advantage actor-critic (A3C). Overflow queries are optimized to crawl the ranked deep web. A3C assigns the reward and penalty to the various queries. Queries are derived from the domain-based taxonomy that helps to fill the search forms. Overflow queries are the collection of queries that match with more than k number of results and only top k matched results are retrieved. Low ranked documents beyond k results are not accessible and lead to low coverage. Overflow queries are optimized to convert into non-overflow queries based on the proposed technique and lead to more coverage. As of yet, no research work has been explored by using A3C with taxonomy in the domain of ranked deep web. The experimental results show that the proposed technique outperforms the three other techniques (i.e., document frequency, random query, and high frequency) in terms of average improvement metric by 26%, 69%, and 92%, respectively.

Health technologies have been and shall always be an integral part of the health system. Appropriate technologies provide solutions to improve healthcare services at an affordable cost. New biomedical, bioengineering and digital technologies continue to swamp the health system and consume a major part of the health budget. National authorities should develop a policy framework that articulates needs, standards and projections of safe and cost-effective technologies in the context of local epidemiological data and the felt needs of communities. Efficient implementation of health technologies requires availability of an adequate number of skilled human resources for health and infrastructure, for maintenance and replacement or for upgradation of these technologies.

The COVID-19 pandemic is a global health and economic crisis of a scale never witnessed before. Beginning in China, it has within a few months devastated many countries around the globe, requiring an unprecedented mobilisation of health systems. While the disease caused by this novel virus is generally mild and self-limiting, the risk of severe disease is disproportionately high among elderly and those with underlying medical conditions. In the absence of a vaccine or treatment, the public health strategies include: (a) preventing transmission through early detection and isolation, tracing contacts and quarantining them and implementation of measures such as social distancing and hand hygiene and (b) reducing mortality by augmenting clinical management and shielding the most vulnerable populations in the society. The pandemic is yet another reminder that we need to invest in public health, ramp up national capacities to detect a disease early and respond rapidly to emerging infections, strengthen and respect our national institutions and rely on evidence base for policymaking. It is high time that we paid heed to these lessons and strengthened without any further delay our health system capacity, as epidemics and pandemics of this nature will continue to challenge public health well into the future.

This paper presents an innovative public–private partnership to revive the urban primary health center (UPHC) system of Nagpur City in Maharashtra State, India. From the baseline assessment, the partnership between the Nagpur Municipal Corporation (NMC) and Tata Trusts identified that the low utilization of UPHCs was due to inadequate infrastructure of the health centers, fewer working hours, inadequately trained human resources, and availability of staff only for a short time. A road map consisting of three phases was jointly prepared by NMC and Tata Trusts to improve the quality of service in 26 UPHCs, of which phase 3 is currently underway.

It describes malnutrition’s impact on health and nonhealth sectors, identifies key determinants, and offers specific solutions according to the local contexts in various urban areas. The recommendations go beyond examining the health sector and take into account water supply, sanitation, sociocultural factors, food supply, and other issues affecting urban nutrition in India.

Context: Reinforcement learning (RL) can help in solving various challenges of deep web crawling. Deep web content can be accessed by filling the search forms rather than hyperlinks. Understanding the search form and proper selection of queries are necessary steps to retrieve the deep web content successfully. Thus, crawling the deep web is a very challenging task. The reinforcement learning-based technique helps in filling the search form and retrieving the deep web content successfully. RL selects the action based on the given state, and the environment assigns reward/penalty to the selected action. Objective: This study reports a survey of RL-based techniques applied in the domain of deep web crawling. Method: Existing literature survey is based on 31 articles from 77 articles published in various reputed journals, conferences, and workshops. Results: Challenges related to various crawling steps of deep web crawling are presented. RL-based techniques are being used in multiple research papers, which solves deep web crawling challenges. Comparative analysis of RL techniques used in deep web crawling is done based on the strength, metrics, dataset, and research gaps. Conclusion: Various RL-based techniques can be applied to deep web crawling, which has not been explored yet. Open challenges and research directions are also recommended.

Background: High inflammatory bowel disease (IBD) rates have been reported in Australasia, but no state-wide studies have yet been performed. Aims: This study estimates the one-year incidence and point prevalence of IBD in the state of Tasmania, Australia. It also reports clinical outcomes after twelve months of diagnosis in an incident cohort. Methods: A prospective, population-based study was performed collecting prevalent and incident state-wide cases from 1st June 2013 to 31st May 2014. Case data were identified from specialist doctors, pathology databases and hospital records. Age-standardised rates (ASRs) were calculated based on World Health Organization 2000 standard population characteristics. Incident cases were followed up twelve months after diagnosis. Results: There were 1719 prevalent cases: ASRs for IBD, Crohn's disease (CD), ulcerative colitis (UC) and inflammatory bowel disease unclassified (IBDU) prevalence were 303.9, 165.5, 131.4 and 6.9 per 100,000 respectively. Prevalent CD cases were younger, with greater immunomodulator/biologic use and bowel resections. There were 149 incident cases: ASRs for IBD, CD, UC and IBDU incidence were 29.5, 15.4, 12.4, and 1.7 per 100,000 respectively. Incident CD cases were more likely than UC or IBDU to require escalation of medical therapy, hospitalisation, and bowel resection especially among those with penetrating or stricturing disease. They had longer duration of symptoms prior to diagnosis. Conclusions: IBD prevalence and incidence rates are high in Tasmania, comparable to data from other Australasian studies and those from Northern Europe and America. Poorer twelve month clinical outcomes occurred in complicated CD, with greater use of healthcare resources.

The necessity for scholarly knowledge mining and management has grown significantly as academic literature and its linkages to authors produce enormously. Information extraction, ontology matching, and accessing academic components with relations have become more critical than ever. Therefore, with the advancement of scientific literature, scholarly knowledge graphs have become critical to various applications where semantics can impart meanings to concepts. The objective of study is to report a literature review regarding knowledge graph construction, refinement and utilization in scholarly domain. Based on scholarly literature, the study presents a complete assessment of current state-of-the-art techniques. We presented an analytical methodology to investigate the existing status of scholarly knowledge graphs (SKG) by structuring scholarly communication. This review paper investigates the field of applying machine learning, rule-based learning, and natural language processing tools and approaches to construct SKG. It further presents the review of knowledge graph utilization and refinement to provide a view of current research efforts. In addition, we offer existing applications and challenges across the board in construction, refinement and utilization collectively. This research will help to identify frontier trends of SKG which will motivate future researchers to carry forward their work.

This paper discusses the stress caused by activities related to academics, especially the stress experienced by new students. These days academics are one of the most dominant causes of stress in a student. Hence this paper proposes a system that would categorize students based on their stress levels. The system is built by combining the various questionnaires and scoring metrics that are famous worldwide and have been developed and backed by leading researchers.The basic structure of the proposed study is to build a system that presents a questionnaire with a defined scoring metric. Based on the answers received from the students, the system will be able to categorize the student’s stress levels. This research revolves around studying various parameters, psychological methodologies, and metrics that are currently being used to monitor users suffering from academic stress. Both global and local research papers and various publications have been studied to find the right set of assessments for the proposed system. For implementation, a questionnaire is used that asks the user various questions based on various parameters - Firstly, Student Academic Stress Scale (SASS). The scales evaluated are – Affective, Behavioural, Cognitive, and Physiological. Secondly, Stress in University Students (SUS) and lastly, Questionnaire of Academic Stress in Secondary Education (QASSE). The research is conducted on various colleges of UT Chandigarh, and results are discussed in the end.

The World Wide Web is expanding at an exponential rate resulting in a vast amount of information. Therefore, search engines need an automated web page classification system to avoid manual classification. Web page classification plays a vital role in deciding the fate of a Focused web crawler by moving it towards the relevant section of the web. Multiple classifiers can be integrated to create a classifier ensemble for web page classification. The performance of the classifier ensemble depends upon the diversity. Diversity can be achieved either by integrating multiple heterogeneous base learners or providing different training subsets to each base classifier. A heterogeneous classifier ensemble results in better predictive performance than an individual classifier. Stacking ensemble method is proposed in this work, integrating Naive Bayes, Support Vector Machine and Random Forest algorithms, as Stacking method, combines the advantages of individual base classifiers and the two-stage learning process of Stacking ensemble leads to better relevance prediction than individual classifiers. This study aims to improve the performance of the classification system by retrieving Indian academician web pages from foreign universities’ websites. The feature set for classification consists of URL text and web page content. Term Frequency-Inverse Document Frequency vectorization assigns feature weights to determine the relevancy of a web page. This work compares the performance of the proposed method with individual classifiers.

In the digital world, World Wide Web magnitude is expanding very promptly. Now a day, a rising number of data-centric websites require a mechanism to crawl the information. The information accessible through hyperlinks can easily be retrieved with general-purpose search engines. A massive chunk of the structured information is invisible behind the search forms. Such immense information is recognized as the deep web and has structured information as compared to the surface web. Crawling the content of deep web is very challenging and requires filling the search forms with suitable queries. This paper proposes an innovative technique using an Asynchronous Advantage Actor-Critic (A3C) to explore the unidentified deep web pages. It is based on the policy gradient deep reinforcement learning technique that parameterizes the policy and value function based on the reward system. A3C has one coordinator and various agents. These agents learn from different environments, update the local gradients to a coordinator, and produce a more stable system. The proposed technique has been validated with Open Directory Project (ODP). The experimental outcome shows that the proposed technique outperforms most of the prevailing techniques based on various metrics such as average precision-recall, average harvest rate, and coverage ratio.

Blockchain technology is reaching mainstream users due to its adoption by a large number of enterprises. Numerous new blockchain platforms and projects are emerging every year, intending to become the best blockchain. However, blockchain technology is far from reaching its full potential because of siloed blockchain platforms. Hence different blockchains should be able to interact with each other for carrying out data sharing and asset transfer among them. In this paper, we propose an approach for atomic cross-chain transfer among heterogeneous blockchains. We provide the implementation details and working of the proposed approach for different Ethereum blockchains.