81,54 €
This book informs the reader about applications of ArtificialIntelligence (AI) and nature-inspired algorithms in different situations. Eachchapter in this book is written by topic experts on AI, nature-inspired algorithmsand data science. Chapters are structured to provide an introduction to the topic, the computational methodology required for experimentation and analysis, and a discussion of results results The basic concepts relevant to these topics are explained, includingevolutionary computing (EC), artificial neural networks (ANN), swarmintelligence (SI), and fuzzy systems (FS). Additionally, the book also covers optimizationalgorithms for data analysis. The contents include algorithms that can be used in systems designedfor plant science research, environmental analysis, computer vision andhealthcare. There are a variety of use cases highlighted in the book that demonstrate how computer algorithms can be used to simulate and understand natural phenomena, such as moving object detection, COVID-19 detection, genetic diversity, physiology and much more. Additionally, the contributors provide useful tips specific to some algorithsm such as load balancing techniques and fuzzy PID controls. The goal of the book is to equip the reader – students and data analysts– with the information needed to apply basic AI algorithms to resolve actualproblems encountered in a professional environment.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 515
Veröffentlichungsjahr: 2003
This is an agreement between you and Bentham Science Publishers Ltd. Please read this License Agreement carefully before using the ebook/echapter/ejournal (“Work”). Your use of the Work constitutes your agreement to the terms and conditions set forth in this License Agreement. If you do not agree to these terms and conditions then you should not use the Work.
Bentham Science Publishers agrees to grant you a non-exclusive, non-transferable limited license to use the Work subject to and in accordance with the following terms and conditions. This License Agreement is for non-library, personal use only. For a library / institutional / multi user license in respect of the Work, please contact: [email protected].
Bentham Science Publishers does not guarantee that the information in the Work is error-free, or warrant that it will meet your requirements or that access to the Work will be uninterrupted or error-free. The Work is provided "as is" without warranty of any kind, either express or implied or statutory, including, without limitation, implied warranties of merchantability and fitness for a particular purpose. The entire risk as to the results and performance of the Work is assumed by you. No responsibility is assumed by Bentham Science Publishers, its staff, editors and/or authors for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products instruction, advertisements or ideas contained in the Work.
In no event will Bentham Science Publishers, its staff, editors and/or authors, be liable for any damages, including, without limitation, special, incidental and/or consequential damages and/or damages for lost data and/or profits arising out of (whether directly or indirectly) the use or inability to use the Work. The entire liability of Bentham Science Publishers shall be limited to the amount actually paid by you for the Work.
Bentham Science Publishers Pte. Ltd. 80 Robinson Road #02-00 Singapore 068898 Singapore Email: [email protected]
This book is based on Applications of Artificial Intelligence and Nature Inspired Algorithms in different areas of Computer Science. Artificial Intelligence (AI) encompasses by means of computers to do things that customarily need human intelligence. It also includes acting on data, learning from new data, and improving over time, just corresponding to a small human kid growing up into a smarter human adult. Nature-inspired algorithms are a set of original problem-solving practices and approaches and take enticing substantial consideration for their respectable act. Typical examples of nature-inspired algorithms contain evolutionary computing (EC), artificial neural networks (ANN), swarm intelligence (SI), and fuzzy systems (FS) and they have been useful to resolve several actual problems. Even with the fame of nature-inspired algorithms, several tests endure, which need extra research efforts. In this book, we focus on Artificial Intelligence (AI) and optimization algorithms for data analytical processes. Each chapter in this book is written by topic experts on applications of Artificial Intelligence (AI) and nature-inspired algorithms in data science.
There has been a tremendous revolution in computing technologies to handle the vast amount of data in recent years. Big data is the large-scale complex data in which real-time data is available and mushrooms the development of almost every field. In recent years, the demand and requirement of big data produced an opportunity to replace traditional data techniques due to their low efficiency and low accuracy. It shows adequate responsiveness, absence of versatility, execution, and precision for meeting the convolution of Big Data challenges. As an outcome, this created different dispersions and innovations. Big data does not mean that the data is humongous but additionally excessive in range and speed. This factor makes them tough to deal with the usage of conventional gear and techniques. Decision-makers read the extension and expansion of big data to understand and extract valuable information from rapidly varying data using big data analytics. In this chapter, we can analyze big data tools and techniques useful for big data. This chapter presents a literature survey covering various applications and technologies that play an indispensable role in offering new solutions dealing with large-scale, high-dimensional data. By summarizing different available technologies in one place from 2011 to 2019, it covers highly ranked international publications. Further, it extends in the context of computing challenges faced by significant Data Healthcare, Clinical Research, E-Commerce, Cloud Computing, Fog computing, Parallel Computing, Pervasive Computing, Reconfigurable Computing, Green Computing, Embedded Computing, Blockchain, Digital Image Processing and IoT and Computing Technology. The survey summarizes the large-scale data computing solutions that help in directing future research in a proper direction. This chapter shows that the popularity of data computing technology has steeply risen in the year 2015, and before 2011, the core research was more popular.
Data refers to bits or chunks of information represented in a digital format in the computational world. This data is available in multiple forms, such as text, symbols, numbers, etc., that are formatted peculiarly (refer to Fig. 1). It is considered an element for modelling or representing real-world events, such as a line segment representing the location of a road. Data represents a binary format for interpreting the process and converting computer data into meaningful information during its transfer. According to Michener, the concept of meta-informatics describes the context, content, structure, accessibility, and quality aspects of the data [1]. Computing covers the activities that require a computer to perform any task, such as processing, managing, transferring, or communicating a piece of information. Here, the concept of hardware and software came into existence.
Fig. (1)) Data Elements.With the emergence of the latest revolutionized tools and technologies, Darmont described the advancement of data-centric computing compared to earlier times [2]. Data analytics and large-scale data mining attracted the interest of various researchers, which led to the existence of a cost-effective solution for online and offline data storage and manipulation, as shown by Sandor et al., while addressing mobile clouds. They proposed “Multi-Authority-based Encryption” to solve the “Key Escrow Problem” in authority-based encryption [3]. From cloud technology and online computing solutions to data storage techniques, all these inherit the risk of threats and attacks. The necessity of data security and privacy has also grown on these grounds. In the last few decades, data has risen to many folds in each field, whether it is dealing with Online Social Media [4], E-Commerce [5], Internet-of-Things [6], Cloud, Healthcare, Bio-Research, or Clinical Data [7]. These big data analytics are characterized by proper data management strategies with supporting architecture to offer better visualization, user interaction, and the development of models. The present review provides a survey of the approaches to deal with the enormously growing demand for a higher level of data computing. Highly cited research articles and relevant research studies are accessed and mined during the process. The work covers the literature supporting applications and approaches based on various technologies to deal with vivid kinds of massive data.
The survey is divided into five sections, including an introduction. Section 2 discusses the survey strategies and provides the application-wise literature review about various approaches; Section 3 provides a descriptive analysis of the work, and Section 4 enumerates the data computing challenges. Finally, the paper summarizes and concludes in Section 5.
The study design involves intensive data mining, which is related to research articles published from 2011 to date in journals and indexed in PubMed, Science Direct, IEEE, Google Scholar, Scopus, SCI with reputed publishers such as Springer, Elsevier, Taylor & Francis, etc. The keywords like “Data Computing”, “Big Data”, “Massive Data”, “Healthcare”, “E-Commerce”, “Cloud”, “Image Processing”, “Biomedical”, “IoT”, “Clinical” are useful with and without “applications”, “tools”, technologies”, “solution”, “approaches”, “analytics” and “server”. The section includes Social Networking, Biomedical Research, E-Commerce [8], Internet-of-Things [9], Online Media, NGS Technologies [10], and Education Sector [11], which is continuously adding up to the data volume and variety. This has irresistibly flooded the network with organized and unorganized data types. Here, the computing framework has revolutionized various data mining, analysis, and visualization strategies. It has also reduced manual efforts. Cloud environment strengthened by fog and parallel computing offers an interesting platform to deal with large-scale multifaceted computing challenges. The technology eradicates the necessity of maintaining the costly computing hardware required to perform various tasks. Moreover, it also effectively addresses software and space requirements. Recent time has seen a tremendous increase in the big data technologies touching the medical and informatics field. We can use data mining by selecting the most recent and relevant research papers in this area. The paper is selected based on variously defined approaches:
Published papers must be in the English language.The publication year should lie between 2011 and 2019.The paper should discuss the design, approach, or application to deal with extensive data.The paper should evaluate the performance of data computing applications.Initially, 160 papers were mined between 2011 and 2019. Furthermore, while employing various inclusion and exclusion criteria, 90 papers were selected. Finally, 73 papers were selected by reading and analyzing the paper content for the study. This is followed by comparing the literary work that is further analyzed regarding data computing applications.
The advent of data-based computing architectures has added pace to the research aspects and resulted in a large-scale data generation with extraordinarily high dimensions and pace. Here, the data is referred to as “Big Data” which is featured by the large volume, veracity, variety, velocity, and value representing larger data sets with complex, massive, and diverse data [12]. The big data revenue forecast published by Shanhong in Statistics 2019 (refer to Fig. 2) predicts that at the end of 2027, the size of the market will hike up to103 billion dollars, which is nearly double the market size of 49 billion dollars in 2019 [13].
A comprehensive survey conducted by Oussous et al. summarized the various aspects of big data computing technology to deal with enormously rising structured and unstructured data with high-tech designs. They also compared big data technology according to their distinct layers like “Data Storage Layer”, “Data Processing Layer”, “Data Querying Layer”, “Data Access Layer” and “Management Layer” [14]. Usha and Aps observed that in the last few decades, Apache Hadoop and Map-Reduce have emerged with the instrumental capability to run a query on terabyte sized data in a couple of seconds that traditional query technologies process in minutes [15]. In this regard, HBase was developed by George that is strengthened by Hadoop technology for high-speed queries from billions of records [16], while Hydra was developed by Lewis et al. as a search engine based on the distributed computing architecture of Hadoop [17].
Fig. (2)) Worldwide Big Data Market Revenue Prediction [13].Cloud offers a robust computational environment and storage space for outsourcing data. The flooding of large-scale data on the network has led to the channelization of data to various cloud servers. These days, cloud storage has been popularly used by both individuals and enterprises. When an individual is sharing data in a public cloud environment, he or she should check for privacy and security aspects. Myrna was developed by Langmead et al. for large-scale cloud computing. Myrna is a tool based on a cloud computing environment used to determine genes in a large dataset of RNA-sequences [18]. Nellore et al. address the privacy protection of cloud data with “Rail-dbGap (Database of Genotypes and Phenotypes)” [19]. “CloVR (Cloud Virtual Machine)” was developed by Angiuoli et al. to address cloud computing environments from the desktop. It is used to manage the data transfer between the desktop and cloud servers. It also improves the performance for processing large datasets using cloud computing resources [20], and a high-performance query system, “Hadoop GIS” was developed using Map Reduce by Wang et al. [21].
Pervasive computing, also considered ubiquitous computing, aims to create an environment packed with all kinds of built-in and portable computing devices that communicate with users’ wireless technologies as they move through this environment. As early as 2006, Shuxin had presented a trust model to deal with the pervasive computing environment [22]. Authors established pervasive computing for creating a revolution by providing services that satisfy user expectations and desires using the information of nearby physical sites. Furthermore, Youssef (2013) has summarized the practicability of pervasive computing regarding cloud services and their adjoining applications. It was found that the cloud environment presents effective communication through devices such as PDAs, mobile phones, and wireless sensors to offer vivid services and facilities [23]. One of the most important features of pervasive computing systems over traditional computing is that it hides out the complex computing functions from the users who are unaware of the computing environment.
Reconfigurable computing is a structural design intended to fill the gap between hardware and software. Koren et al. (2018) offer potentially much higher performance than software while retaining a greater degree of flexibility of hardware components [24]. Earlier in 2002, the author had described Reconfigurable devices, including Field-Programmable Door Arrays (FPGAs). There are several computational elements that are defined by the programmable configuration bits. Campton et al. examined that these elements, often known as logic blocks, are linked by means of a collection of programmable routing resources [25].
Green Computing refers to the efficient use of computers and other technologies while respecting the environment through energy-efficient peripherals, environmental protection, and reducing electronic waste. Mishra et al. (2018) demonstrated that these goals not only make resources more effective but, at the same time, enhance the overall performance of the system [26]. Naji et al. (2017) summarised that green computing had been described in two ways [27]:
Software Technology: The main goal is to create techniques to improve software, storage, and energy efficiency.Hardware Technology: The goal is to reduce energy consumption without presenting expertise that can create economically efficient technology assisted with the concept of recycling.“Embedded System” is an acronym representing “Embedded Computer System” that embodies the system product or electronic device in which a computer has been incorporated. Furuichi et al. (2014) presented “Internal system” as a general term for any electronic device that includes any system, product, and computer [28]. Jakoyljevis examined the “Embedded Systems” consisting of a series of software and hardware modules designed to support and sustain manageable and certifiable integration and hosting of various control and maintenance functions [29].
It is one of the advanced computing technologies that have recently been implemented to deal with enormous data volumes. CloudBurst is one of the parallelized computing models that facilitate the genome mapping process developed by Schatz, and for “Short-read mapping” process. It uses multiple nodes to parallelize the execution of processes using the “Hadoop Map Reduce” framework [30]. Keckler et al. (2011) detailed the salient features of implementing a high-throughput architecture that is based on GPU. They also considered the parallel computing challenge with its potential effects on the research community [31]. Zhang et al. (2019) postulated various parallel computing-based solutions to deal with the modelling of multidimensional statistical analysis of the Markov Chain Random Field model. In the methodology, Multicore Processor Parallel Computing, GPU-based Nearest Neighbour Searching (GNNS) Parallel Computing, and Multicore Processor-based GNNS were also used. Simulation analysis employed both CPU and GPU during the execution of computing solutions. The results demonstrated that the proposed parallel computing designs were 1.8 units faster [32].
Many applications are available that are hybrid approaches involving various technologies together to improve the technology outcomes. Tortonesi et al. (2019) proposed a “new information-centric” and “Value-based service” model designed as “SPF Fog-as-a- Service Platform” which can be used to tackle the important challenges in a smart city i.e., to process a massive amount of raw data generated in a smart city. They established that fog computing is a wiser combination of WSN, IoT, mobile, and edge computing technology. The concept proved to be well suitable for smart city concepts. Various applications related to business and civil service can also be run on “Fog Computing” platform. The authors described how fog computing governs the location, time-sensitive, and context-based applications covering sensitive information of both consumers and producers to offer a reasonably cost-effective platform [33]. For instance, Fog computing architecture is an enhanced cloud computing platform combining designs for enhancing IoT based applications in the industry. This proposed work by Jang et al. (2019) summarized the design of smart factories for the future. The technology-driven intelligent, self-monitoring and self-organizing design proved to be nearly autonomous to adapt to the factory's manufacturing design. The system was successfully evaluated against a permission-based fog computing system that was inspired by decentralized technology [34]. Communication and computation are the essential aspects of Fog Computing. The rush of vehicles increases in downtown during working hours and most vehicles are parked in the parking areas and on the roadside for a long time. Thus, these movable and immovable transports can be used as resources for communication and computing. Ergo Hou et al. proposed a “Vehicular Fog Computing” in which they used vehicles as an infrastructure to forge the better usage of these automobiles’ conveyance and calculations [35]. Laun also considers Fog computing as the best solution for providing exemplary mobile services to users at any place with proper usage of resources and for better service of mobile traffic when they use location-based services of the mobile devices [36].
Internet-of-Things (IoT) has also shown wide data computing applications developed across wider domains, including Production, Occupation, Shipping, Health Protection, and Households. For instance, Breidenbacher et al. presented HADES-IoT in the year 2019 as a “Tamper-Proof” host-based anomaly detection design for IoT devices to protect them from malware devices. They offer the last line defence mechanism for installing a range of IoT devices advantaged with Linux platform. The design is 100% effective in detecting IoT, Reaper, and VPN Filters while utilizing nearly 5% of the memory [37]. Yun designed one M2M as a standardised IoT platform for assisting the sensing capabilities of one M2M-defined web-based application. The design successfully dealt with the heterogeneous hardware interfaces of IoT devices [38]. Dimitrov (2016) focussed their study on various MIoT “(Medical Internet of Things)” devices and big data for computing data related to healthcare facilities offered by hospitals and various government and private organizations. A new category of “Digital Health Advisors “will help their clients avoid illness related to their diet, improve their mental health, and achieve a better lifestyle. There are so many devices and mobile apps like “Myo”, “Zio patch”, “Glaxo”, “Novartis” that are used for smart healthcare. People have started using mobile apps and IoT devices to monitor their health, remember their appointments, calculate the calories they burnt while walking or running, and easily communicate with their doctors [39]. Bhattacharyya et al. (2018) designed an exciting and interactive gaming app in Visual Studio (IDE) using user’s “myoelectric signal” to aid in the rehabilitation of patients recovering from heart diseases, stroke, depression, surgery and design and control of orthotics, prosthetics, etc. It offers an “MYO armband” that needs to be tied to the recovering patient to recognize various gestures of hand movements and uses the in-built “EMG (Electromyography)” electrode for data retrieval [40]. Liu et al. (2018) dedicated their research to designing IoT based “patch-type ECG monitor “applications to offer real-time monitoring apps and used to measure the heartbeat of patients fighting with “cardiovascular disorders” [41]. The rising demand for encryption in every field has led to the emergence of novel ideas to deal with privacy and security aspects. Singh et al. (2019) proposed an IoT-based smart home design using decentralization technology of blockchain and cloud computing. To achieve confidentiality and integrity in block chain technology, they use encryption and hashing algorithms. Smart home security aspects were dealt with while monitoring the network traffic to establish a correlation between network traffic and traffic features so that they can use the “Multivariate Correlation Analysis (MCA)” identification technique to measure the flow of smart network traffic. This helps to categories the association between the features of traffic [42]. The blockchains offer the most efficient solution to IoT based design.
Blockchain offers a kind of one-time design and creation of databases that can only be read without any scope of removal or modification. It offers decentralized data accessible by the owner with the help of private keys. Kaye et al. (2015) designed a dynamic patient consent system that worked as a personalized communication interface. This interface helps user-centred decision making, where participants can securely share their data with third parties in an encrypted form and offers a single platform for intermingling their interests and concerns [43]. Blockchain inspired design offers encryption of sensitive data when the records and samples are shared. Globally, blockchain has shown potential applications in maintaining Electronic Medical Records to aid in the apt functioning of the Food and Drug Administration. Cyran (2018) had designed a blockchain-based solution that was inspired by the microservices framework [44].
The design offered an independently secured encapsulation of different services customizable to meet individual hospital needs. The core design covered interaction among nodes, cryptographic security, extensive data file sharing facilities, and integration of logical business sharing using smart contracts [45]. Dimitrov (2019) published his research focussing on the applications inspired by blockchain architecture. He was mainly concerned about offering solutions to data management issues faced by healthcare centres [46]. Field exhibiting tremendous data computing applications are shown in Fig. (3) and Table 1.
Fig. (3)) Data Computing Applications in Various Fields.Schuster published an article in which he established that the rising interest in the new generation of “Non-Sanger-based Sequencing” technology, i.e., Next-Generation Sequencing, transformed the traditional design of genomic and expression analysis [47]. For instance, NGS technology could process billions of DNA sequence data in a day and gives maximum throughput than existing methods, as demonstrated by the study published by Knetsch et al. in 2019 [48]. Recently, in 2019, Seth et al. had also discussed the big data tools and technologies that majorly focussed on effectively addressing issues related to data storage, data retrieval, data analysis, data integration in the vivid platform, and detecting elements raising errors. Security has always been the main concern while storing data on the cloud. Hence, they proposed a mechanism for efficient storage of data using fog computing as an extended version of cloud computing [49]. Critical analysis of various achievements and hurdles for the transformation from Landscape Genetics” to “Landscape Genomics” was presented by Cushman et al. in his published study focussing on NGS data [50].
The concept of Data Computation has an inevitable role in digital image processing applications in vivid fields like biometrics and medicine. For instance, in 2017, Ghazali et al. had addressed the problem of colour image processing with specific applications to weed classification [51]. In this work, they achieved higher classification accuracy by combining both colour and texture information of weed images. For preprocessing, a task colour filtering algorithm was used and the proposed technique and known as the extracted green colour. The two feature extraction algorithms, Gray level co-occurrence matrix (GLCM) and Fast Fourier Transform (FFT), were used for classification. Colour-based method is better than the gray scale for better classification results. Hassan et al. (2019) described novel methods used for detecting and classifying the external defects of olive fruits based on texture analysis and texture homogeneity [52]. The proposed technique distinguished the defected and healthy olive fruits and was further used to identify and recognize the real defected region. The outcomes were compared with image processing techniques such as Canny, Otsu, local binary pattern algorithm, K-means, and Fuzzy C-means algorithms, demonstrating the highest accuracy achieved by the proposed technique. Mia et al. (2018) proposed an efficient approach based on Linear Discriminant Analysis (LDA) for initial image segmentation by an unsupervised method [53]. The proposed method was capable of automatically separating and combining the clusters whenever required. The proposed method outperforms various cluster-based image segmentation methods, such as the k -means algorithm, the SOFM and LDA segmentation method, and the LDA and K-means based segmentation. The authors established that the performance of the proposed algorithm can be improved by using machine learning algorithms. These are important for the segmentation of natural images and medical images. In 2019, Trivedi et al. discussed image informatics as more related to the various aspects of medical image processing used in relation to “Digital Imaging and Communications in Medicine (DICOM)” format images, including tissue scans [54]. The concept for improvement of the quality of medical diagnosis has raised the interest in computer-based biomedical image processing and analysis. Radiology imaging provides depth visualization of various complex biological systems, tissues, and organs. In this area of research, Saba et al. had implemented Deep Learning (DL) Technology to analyse radiology data. With the proper use of technology, the value of the radiologist is enhanced with the dormant of DL in the conveyance of healthcare. It was conducted by upgrading the patient’s results with minimal cost. They also discussed three fundamental challenges viz. “Safety”, “Privacy”, “Legality” in the diagnosis of radiology data [55]. Image diagnostics are widely available in the form of MRI, CT and X-Ray scans performed to distinguish benign and diseased tissues visually. The process is governed by implementing various Artificial Intelligence approaches to create intelligent systems to mimic the human brain. To achieve this, various Machine Learning approaches powered by image optimization strategies are implemented. For instance, Higaki et al. used Deep Learning Architecture for improving the quality of images in different expertise as “noise and artifact reduction”, “super-resolution” and “image acquisition and reconstruction”. They developed a deep learning-based “denoising convolutional neural network” to remove noise from the image [56]. Gyftopoulos et al. used artificial intelligence [57] and DeSouza et al. used multimodality imaging techniques [58]. Here, another important aspect is intelligent data selection or region selection from the whole image scan.
Kagadis et al. (2013) discussed cloud computing as a practical solution that offers resource reconfiguration of platforms, applications, and virtual environments in a very cost-effective manner. The researchers and clinicians have been inspired by the vital role being played by cloud computing and cloud storage facilities. The authors have majorly focussed on biomedical imaging technology, taking advantage of computing platforms [59]. Ali et al. (2018) described innovative designs to offer technology-driven services to the healthcare sector. The work involved a rigorous evaluation of the cloud data computing environment to cover applications, issues and offered opportunities for the emergence of cloud computing in healthcare. The study showed that cloud computing enhanced the quality that led to decision-making, security, and privacy [60].
Various E-commerce websites are also constantly generating a large volume of business data. The enterprises also extend their services to customers by learning individuals' interests. To avoid wasteful data searching, opinion and sentimental mining had shown great prospects. Kansal and Kumar have proposed financial sentiment analysis on social network data from financial markets, Facebook, Twitter, and news using artificial intelligence and Cuckoo search [61]. Derindag et al. (2019) presented an analytical study covering the technologies to address e-governance and e-commerce. They also critically analysed the impact of the revolution on the economic sector with various growth indicators [62]. Kansal and Kumar have also developed a stock market forecasting system that demonstrated 94.44% accuracy for emotion-based text categorization. The designed stock price forecasting system has implemented a cuckoo search with neural networks to obtain instrumental prediction results [63].
The recent revolution in technologies has added pace to the research aspects of healthcare informatics, biomedical data analysis and resulted in large-scale data generation related to biological research with extraordinarily high dimensions and pace. The current section moves around analytics, majorly covering biomedical research about clinical informatics, including diagnostics. Chandana et al. in 2018 defined that healthcare informatics addresses the tools and techniques that are used to estimate and offer solutions to the breakout of the most hazardous infectious diseases. The proposed combined model “Convolutional Neural Network” using “Naive Bayes”, “Random Forest” and “Linear Regression” on “Hadoop platform” to forecast lung diseases at the beginning stage of the health service group [64]. Tavazzi in 2019 stated that clinical informatics deals with clinical trials and intelligent decision making to benefit patients in the light of knowledge and evidence [65]. The rising trends of personalized medicine have produced massive data covering large data sets. Mass patient selection, designing protocols, mass evaluation, and conducting various levels of clinical trials has become feasible with computing technologies. Therefore, storing an amount of electronic medical records efficiently is itself a great challenge. Software solutions like MedCloud based on HIPAA [66] and Cloudwave [67] were developed by Sobhy et al. and Jaypndian et al. to manage clinical data. Online data management has resulted in easy channelization of information that needs to be shared with patients, volunteers, drug industry, pharmacists, doctors, and professionals. Nipp et al. figured out various facts to enlighten individuals with many misconceptions and thereby encouraging higher patient recruitment under various trials [68].
The review has shown that data computing has not only explored informatics related fields like cloud, IoT, social networks, enterprise, and industrial networks. Further encouragement proceeds with potential computing solutions to answer large-scale data computing issues of the biological world dealing with big genomic data, practical aspects of medical research, disease prognosis, and diagnosis. The strategies like patient participation, enrolment, and clinical trials have also been positively changed with the help of online data management assays with the help of blockchain-based secure cloud storage. Fig. (4) summarizes the relationships among various computing technologies along with their applications in various fields. It has also been observed that the various data computing technologies cannot be isolated, and to harvest the best computing results, secure storage space and communication channels are an essential part of data computing approaches.
The year-wise distribution of the publications considered for conducting the present study is shown in Fig. (5). It is observed from the pie chart that very few instances for applications of data computing are seen from 2011 to 2017, which corresponds to 8% to 15%. It is also observed that 57% of the total publications used in the survey have been published between 2017 and 2019. This means that data computing has been a recent technology that has shown its impact in dealing with larger volumes of diverse data. Overall, it can also be summarized that 75% of the survey covers the data computing applications that were published not before 2013. Hence, it can be said that from the year 2017, a new era of revolutionized data computing technology has started.
Fig. (4)) Data Computing Architecture. Fig. (5)) Publications Referred in the Survey.Application wise distribution of publications covered in the current survey is shown in Fig. (6). The data computing applications from 2011 to 2019 are plotted on the Y-axis with the number of publications used in surveys on the Y-axis. It is observed that maximum data computing applications deal with large-scale data applications, namely, IoT and Big Data analytics, followed by cloud computing, computational informatics, Digital Image Processing, healthcare and Clinical Research. The trend is further followed by the latest emerging technologies like blockchain, NGS, e-commerce, parallel computing, and fog computing that may surpass the popularity of the existing data computing applications in the coming years.
Fig. (6)) Data Computing Applications from 2011 to 2019.Revolutionizing data computing technologies is constantly raising challenges regarding the variety and volume of data and that too at a very fast pace. This has subsequently challenged various aspects such as data availability, scalability, quality, security, privacy aspects, regulatory policies, and governance.
Highly efficient large-scale data management strategies are required to deal with scalability concerns. In this regard, G-store was introduced by Das et al. offering key based structured data storage and retrieval with easy API [69].Data availability summarises the apt inputs, or the raw data should be constantly available as a resource for data computing technologies. Demand for a constant and high-quality resource needs to be addressed [70].Data integrity is another aspect of data resources that means that the data should remain unchanged or modified in any form during transfer, storage, or analysis. This aspect has been addressed with the standardisation of medical imaging in terms of DICOM and much more needs to be done in numerous fields [71].Large-scale computing usually results in unstructured or semistructured data. This data needs to be converted to structured data to aid further analysis. The challenge here lies in the data transformation strategies, as they should not alter data during the transformation process [72].Data security and privacy are indispensable needs of the current large-scale computing architectures. Data over the network or cloud environment is highly prone to theft and malicious attacks, exposing sensitive data fields to stealing, profiling, and loss of control. Vivid encryption strategies are available, including privacy preservation in the cloud, still, more customized security solutions are in high demand [73].Laws, rules, and regulations should safeguard sensitive data from unauthorised use. Adequate policies need to be developed to offer protection mechanisms and protect the enormously rising data volumes. Data governing policies, including monitoring, and monetizing information, should be developed to benefit the consumer and business sectors.The nature of these challenges is constantly changing with the revolutionizing computing trends and technology advancements that need to be wisely addressed to harness the full advantage of the latest applications.
The authors provided a newly created intelligent approach for assessing the IoT smart system's dependability. A machine learning approach known as a decision tree is used to validate the data of the smart meter. The smart meter readings were regressed and classified using the decision tree approach into real and false types. In addition, the proposed system can identify data loss caused by erratic internet signals. The IoT platform's dashboard displays the system's online output, such as data loss, genuine and false data. Three scenarios have been used to assess the efficacy of the planned infrastructure. Scenario 1 demonstrates that the data is true and that there is no loss, and that the decision tree model is error-free. In scenario 2, it is demonstrated that the decision tree model can detect false data, allowing the user to secure and verify the smart meter. The traffic light has been changed to a red color to hint to the user about the abnormal case of fake data. The suggested infrastructure was tested in the final scenario to check for data loss and publish the findings on the IoT platform's dashboard, where it was determined that the network speed was insufficient for the smart system. In general, the suggested technique improves the dependability of smart IoT systems, which boosts industry 4.0 investment [74].
The survey showed that data volume is constantly rising with the revolutionized data computing technologies, data sensors, and mobile devices connected via the internet, offering diverse functionalities. Impossibility concludes to demarcate the big data from cloud computing and blockchain architecture. Additionally, further advanced data computing has been offered with the involvement of edge, fog, and parallel computing environments. Numerous applications have also been designed that use blockchain architecture to deal with the security and privacy aspects of cloud and IoT applications. In biomedical research, computing technology has revolutionized diagnostics and research protocols. Digital Imaging Processing has also been improved considerably, with data computing and data transfer of attributes related to the healthcare sector. In the education sector, data computing facilities have revolutionized the architecture of online classroom teaching and training based on Hadoop based online classes. Literature review from authenticated publication sources has shown that recent years from 2015 to 2019 have shown a tremendous rise in the development and deployment of computing solutions in every field. Hence, it can be summarized that data computing is a general term that has touched every aspect of life and is a popular area drawing the attention of the research community. However, there are rising instances of security attacks and privacy concerns that are being successfully addressed by computing technology. In light of the current computing survey, a highly revolutionized computing paradigm can be predicted in the near future.
Not applicable.
The authors declare no conflict of interest, financial or otherwise.
The authors are thankful to the learned reviewers for their valuable comments.