Colossal Data Research Articles

Scalability and Efficiency of Deep Learning Models on High-Performance Computing Clusters: Bibliometric Analysis

Purpose: This systematic review identifies the main advancements, core papers, voids, and outlooks regarding utilising high-performance computing (HPC) clusters for large-scale deep learning models. Moreover, it aims to assess research trends, significant approaches, and methods to improve the effectiveness and adaptability of these models. Design/Methodology/Approach: This paper involves a systematic literature review and bibliometric analysis of published articles from 2018-2023. To limit the results, some specific keywords have been introduced: Web of Science, deep learning efficiency, scalability, and HPC clusters. Studies were screened according to the PRISMA flowchart, covering 364 articles, 19 of which were included in the systematic review based on further criteria. Findings: The bibliometric analysis revealed that the most globally cited articles were from IEEE Transactions on Emerging Topics in Computing and Joule. However, the most relevant sources identified were the Journal of Supercomputing, Concurrency and Computation: Practice and Experience, and IEEE Access. Researchers from the USA, China, Korea, and the UK authored the most significant contributions. Research Limitation: The study examined only works from the Web of Science database from 2018 to 2023. Practical Implication: The proposed research results contribute crucial information to enhance the effectiveness of deep predictive models in large-scale HPC environments, which are essential for enterprises adopting artificial intelligence (AI) and machine learning (ML) methodologies in colossal data analysis applications. Social Implication: Propelling deep learning models with the help of HPC clusters can create more vital AI solutions that can respond to society's needs. Originality/ Value: The novelty of the research stems from the bibliometric assessment and the question of which sources and authors in this field are the most important.

Artificial Intelligence (AI) Importance - A Brainchild of Humans!

Human Brain Intelligence cannot be Dwarfed by AI (Artificial Intelligence), a Brain Child of Human, but a Remarkable Tool to handle Colossal Data more Efficiently with a Greater Speed, Accuracy using Smart Algorithms Devoid of Emotions & New Science!

Managing Cyber Security Debt: Strategies for Identification, Prioritisation, and Mitigation

This paper explores cyber security debt, a technical debt arising from unaddressed security vulnerabilities in an organisation's IT systems. These vulnerabilities accumulate due to resource limitations, time constraints, and expertise gaps, potentially leading to security breaches and data compromises. The paper outlines the cyber security debt management process involving identification, prioritisation, and mitigation strategies. Drawing parallels to financial debt, the authors emphasise the escalating risks of delaying cyber security debt repayment. The paper underscores the significance of diligent debt management in maintaining digital resilience and mitigating cyber threats. The increasing interconnectedness of systems and rapid software development has given rise to a hidden challenge known as cyber security debt. Cyber security debt is posed as a subset of technical debt, encompassing the accumulation of security vulnerabilities within an organisation's IT infrastructure and applications. Drawing a parallel between cyber security debt and its financial counterpart, the authors underscore the grave risks of deferring debt repayment. Just as financial debt accrues interest, unresolved security vulnerabilities compound over time, elevating the likelihood of breaches and data exposure. A poignant case study of the Equifax breach exemplifies the real-world consequences of neglecting security debt management. The failure to patch a well-known vulnerability led to a colossal data breach, highlighting the urgency of addressing security weaknesses promptly. Complex in nature, cyber security debt materialises when organisations fail to address vulnerabilities during various operational life cycles. These vulnerabilities might remain concealed within IT architecture, legacy code, or third-party libraries, posing a formidable challenge to detection and resolution. By understanding the parallels between financial and cyber security debt and proactively managing the latter, organisations can enhance their ability to safeguard against evolving cyber threats and maintain a robust security posture.

A Blockchain-Based Secure Searching Strategy for Metadata in Mobile Edge Computing

With intelligent devices’ prevalence, mobile edge computing technology can effectively process the massive data in the Internet of Things (IoT). The metadata produced in mobile edge computing plays a role in colossal data management. However, it also includes lots of user privacy information. Traditional privacy protection solutions sacrifice the metadata’s availability or are incompatible with the distributed environment. To overcome this dilemma, we propose a blockchain-based secure searching strategy for metadata (BSSMeta) in mobile edge computing. By leveraging a lightweight proxy re-encryption scheme and a decentralized key generation method on the blockchain, BSSMeta keeps the security of the metadata searching tasks. Meanwhile, a master-slave smart contract method and a buffer uploading strategy are proposed for the efficiency of the searching and uploading processes. As a byproduct, BSSMeta also supports searching metadata in a multi-user environment and gives users access control to metadata. Finally, we provide a security analysis of BSSMeta and implement a prototype on the Hyperledger Fabric platform. The results of experiments on various workloads show that BSSMeta is feasible and flexible in security and efficiency.

From insights to impact: leveraging data analytics for data-driven decision-making and productivity in banking sector

This research stems from the disturbing phenomenon known as digital transformation and the colossal data creation. A vast amount of data has been produced as a result of increased usage of digital technologies to generate commercial value. However, data does not hold any value, so organizations are motivated to invest in data analytics and make informed decisions to enhance their performance. Against this backdrop, this study investigates the impact of data-driven decision-making (DDDM) on productivity in the presence of data analytics capability of Pakistan’s banking sector. We explore this link based on innovation diffusion theory using Instrumental Variable Two-Stage Least Square. A composite index of DDDM was developed based on primary data collected through online survey. This index is then regressed on actual productivity measures for 2016–2020. The findings suggest that banks exploiting analytics and adopting DDDM methods results in an increase in productivity of about 9–10%. It further indicates that DDDM adoption with an investment in data analytics leads to enhanced productivity.

A new miniaturized wideband self-isolated two-port MIMO antenna for 5G millimeter-wave applications

Nowadays, millimeter-wave frequencies present a catchy solution to securing the colossal data rate needed for 5G communications. Accordingly, this research deals with the conception of a novel orthogonal 2×2 multiple input, multiple output (MIMO) antenna design operating in the millimeter wave spectrum with quite small dimensions of 11×6×0.8 mm 3 . The single antenna element consists of a trapezoidal microstrip patch antenna built on the Rogers RT5880 laminate with a permittivity of 2.2 and tangent loss of 0.0009. A trapezoidal-slot ground plane is used to support the structure. The antenna resonates at 28 GHz with a large bandwidth of 4 GHz from 26 to 30 GHz, a good gain of up to 5 dB, and a high radiation efficiency of 99%. A strong isolation is achieved that surpasses 26 dB. Besides, a high diversity performance is achieved where the envelope correlation coefficient (ECC) is lower than 0.001, the diversity gain (DG) is greater than 10 dB, and the channel capacity loss (CCL) is no longer than 0.4 bit/s/Hz. The achieved outcomes prove the robustness of the suggested MIMO antenna and qualify it to be a strong candidate for 5G wireless devices.

Building a proper churn prediction model for Vietnam's mobile banking service

This study aims to build a model predicting the churn rate of customers using mobile banking services in Vietnam by applying data mining techniques. Customer churn is an issue that any service provider must pay attention to because it is decisive to the development of the business. The competition between banks is getting tougher, hence customer churn prediction has become of great concern to banking service companies. It is necessary for banks to collect colossal data and establish a valued model for classifying types of customers. In this study, three supervised statistical learning methods which are KNN, Random Forest, and Gradient Boosting are applied to the churn prediction model using the data source of VIB’s customers. In addition to selecting models belonging to the group of weak single learners such as Neural Networks, Naïve Bayes Classifier, and K-nearest Neighbor..., this paper utilizes Random Forest and Gradient Boosting which are assessed as better models because they can combine weak learners for improving model efficiency and capable of classification. The results exhibited that Gradient Boosting is the best performance in the three above classifiers with a 79.71% of accuracy rate, and 86.23% of ROC (Receiver Operating Characteristic) curve graph. Moreover, the decision tree algorithm generates readable rules for churner and non-churner classification which are potentially helpful to managers. Finally, this study suggests a proper model that can be used to forecast churners of mobile banking services in Vietnam.

Review of Semantic Importance and Role of using Ontologies in Web Information Retrieval Techniques

The Web contains an enormous amount of information, which is managed to accumulate, researched, and regularly used by many users. The nature of the Web is multilingual and growing very fast with its diverse nature of data including unstructured or semi-structured data such as Websites, texts, journals, and files. Obtaining critical relevant data from such vast data with its diverse nature has been a monotonous and challenging task. Simple key phrase data gathering systems rely heavily on statistics, resulting in a word incompatibility problem related to a specific word's inescapable semantic and situation variants. As a result, there is an urgent need to arrange such colossal data systematically to find out the relevant information that can be quickly analyzed and fulfill the users' needs in the relevant context. Over the years ontologies are widely used in the semantic Web to contain unorganized information systematic and structured manner. Still, they have also significantly enhanced the efficiency of various information recovery approaches. Ontological information gathering systems recover files focused on the semantic relation of the search request and the searchable information. This paper examines contemporary ontology-based information extraction techniques for texts, interactive media, and multilingual data types. Moreover, the study tried to compare and classify the most significant developments utilized in the search and retrieval techniques and their major disadvantages and benefits.

A Comprehensive Survey for Hadoop Distributed File System

In the last few days, data and the internet have become increasingly growing, occurring in big data. For these problems, there are many software frameworks used to increase the performance of the distributed system. This software is used for available ample data storage. One of the most beneficial software frameworks used to utilize data in distributed systems is Hadoop. This software creates machine clustering and formatting the work between them. Hadoop consists of two major components: Hadoop Distributed File System (HDFS) and Map Reduce (MR). By Hadoop, we can process, count, and distribute each word in a large file and know the number of affecting for each of them. The HDFS is designed to effectively store and transmit colossal data sets to high-bandwidth user applications. The differences between this and other file systems provided are relevant. HDFS is intended for low-cost hardware and is exceptionally tolerant to defects. Thousands of computers in a vast cluster both have directly associated storage functions and user programmers. The resource scales with demand while being cost-effective in all sizes by distributing storage and calculation through numerous servers. Depending on the above characteristics of the HDFS, many researchers worked in this field trying to enhance the performance and efficiency of the addressed file system to be one of the most active cloud systems. This paper offers an adequate study to review the essential investigations as a trend beneficial for researchers wishing to operate in such a system. The basic ideas and features of the investigated experiments were taken into account to have a robust comparison, which simplifies the selection for future researchers in this subject.
 According to many authors, this paper will explain what Hadoop is and its architectures, how it works, and its performance analysis in a distributed systems. In addition, assessing each Writing and compare with each other.

NACO predicated hybrid model of internet of things and cloud computing to manage immensely colossal data in health accommodations applications

NACO predicated hybrid model of internet of things and cloud computing to manage immensely colossal data in health accommodations applications

Remodelling State-Space Prediction With Deep Neural Networks for Probabilistic Load Forecasting

Probabilistic load forecasting (PLF) has become necessary for power system operators to do efficient planning across power transmission and distribution systems. However, there are not many PLF models, and those that exist take a lot of computation time and are not efficient, especially in multiple loads. This paper proposes a novel algorithm for spatially correlated multiple loads wherein a global parameter is learned from state-space parameters of individual loads by an amalgamation of deep neural networks and state-space models. The proposed model employs complex pattern learning capabilities of recurrent neural networks and temporal pattern extraction of innovation state-space models. It is tested on GEFCom-14 and ISO-NE datasets, one with a single load and multiple loads. Different case studies are conducted to examine the involvement of temperature for load forecasting. It has been observed that in the case of multivariate loads, temperature variable doesn’t make much difference in PLF, but in the case of univariate loads, forecasting results are four-times better. The proposed method is highly interpretable and can be employed in areas where limited training data is available to the areas where colossal data is available. The proposed model has outperformed several benchmarks present in the literature on the same datasets.

A PROPOSED MINING APPROACH BASED ON BUSINESS INTELLIGENCE BY USING DATA SCIENCES TECHNIQUES

The progression of the web has caused digitalization of data which opens up colossal data openings. Online data in colossal numbers leaves traces of what customers see, what they read, their affiliation and lead, judgment, about their tendencies and tendencies to give a great deal of data that can be burrowed for learning The estimation of enormous data lies in the results of examination and figures or exercises taken from the eventual outcomes of the assessment and desire from Jumia Website. Prescient examination is information use, measurable calculations, and AI methods to recognize potential patterns, occasions, and practices later on dependent on authentic This paper attempts to propose prescient investigation to foresee Consumer Behavior conduct on Jumia Website in Egypt by utilizing conduct information, informatics and investigation approach by utilizing Data Sciences Techniques through grouping Classification Technique by choice tree Algorithm in particular Random Tree the Results show that this the Random Tree Algorithm is reasonable for Jumia Datasets for anticipating Consumer conduct with the goal that more profound knowledge into Consumer conduct can be gotten to help prescient examination so as to improve business choice

Key-Value Combination for Scholarship Disbursal Analysis through Map-Reduce

Big Data includes both structured and non-structured contents. Big data is being generated by devices and increasing in volumes. The social and digital landscape produces this colossal data, although devices, handheld, and IoT devices add to the size. Data set with respect to a Central University scholarship domain in a country like India can make the dataset still larger and complex. The complex dataset in the coming years is sure to certify as Big data in terms of velocity, volume, variety, veracity, and value, thus various Bigdata analytical tools are needed for analytics of this huge dataset. The purpose of research on such a dataset is to retrieve meaningful information that facilitates decision-makers for quick and flawless decision making and redirecting scholarships to the deserving candidates across the country. This paper explores the storage of big data and map reducing through a key-value analysis for scholarship disbursal decision making thus catalyzing Digital India vision.

Applying attribute-based encryption on mobile devices

The 21st century has witnessed the rapid development of small and convenient mobile devices such as smartphones, tablets, game players, sensor nodes, etc. The rise of such mobile devices indicates the increase of colossal data transmission through the Internet and online services along with the challenges of data security. It is common to think of a solution to protect sensitive data from unauthorized users, and the most popular solution is to use encryption. While many research activities in functional encryption have widely been applied to network devices, computers, and applications, mobile devices still attract much attention to security issues due to the limitations of system resources, connectivity, data transmission and power consumption that malicious users can exploit to launch attacks. Especially, mobile devices have become a principal tool to share data on the Internet through online services, such as Facebook, Youtube, DropBox, Amazon, Online Games, etc. This paper presents a study of the Attribute-Based Encryption (ABE) scheme that exploits user attributes to build the secret key and the ciphertext. ABE encryption is specified by a set of attributes or a policy defining attributes that users possess. The paper also describes a few implementations of ABE applied in the cryptography community and the challenges of integrating ABE into real-world applications. Finally, the paper proposes an implementation of ABE for Android mobile devices. This implementation associated with the Kerberos protocol can be applied to secured data sharing applications. The Kerberos protocol aims at providing mutual authentication for the client server model. Experiments have evaluated the proposed ABE implementation on Android mobile devices along with the Kerberos system. The evaluation also includes ABE performance with discussions and lessons learned.

Fuzzy Object Oriented Database for Big Data

Due to the huge amount of uncertain data collection, mankind facing a colossal amount and fast data having such a composite configurations. The data can be transact on the web and manually exchanges, interpersonal organizations and through our everyday life exercises. A very helpful result of designing an appropriate Big Data in different zones, for example, medical services, human services, executives and administrators. Managing the Big Data efficiently, new investigation perspective has been secured at this point the viewpoints about huge information (Big Data) requires extra-long stretch innovative interests. The Fuzzy Logic s have been implemented here for Big Data because of their capacities to handle the vagueness and uncertainties in the information. A couple of imaginative approach for Big Data processing is presented previously. To abbreviate the current duties and extant a point of view of further enhancements. We survey the different examinations and concentrated on that there are constrained Fuzzy systems have been adopted in large information (Big Data preparing. We also examine the various advantages of Fuzzy Logic s in large information (Big Data) issues. Therefore in this paper a Fuzzy object-oriented database is designed for Big Data and perform some Fuzzy queries to check the performance of the designed Fuzzy object-oriented database. We focused on the continuously propelled augmentations of Fuzzy sets and their blends in with various contraptions could offer a novel promising planning condition.

Methodological Research on Architecture of Cloud Service Models in Cloud Computing

Cloud computing is the most recent association in speaking to computing resources as an assistance. It addresses a move away from figuring as a thing that is gained, to computing as a help that is given to clients over the web from colossal scale data focuses – or "clouds". While cloud computing is expanding making acclaim in the IT business, the academic world appeared to fall behind the speedy movements in this field. In this paper, we investigate cloud building and differences cloud figuring and framework handling. We also address the characteristics and uses of a couple of surely understood cloud computing stages. In this paper, we discuss about the architecture of cloud computing models for which we perceived a couple of troubles from the cloud figuring gathering point of view what's more, we additionally highlighted the cloud interoperability issue that preferences broad future scope & progress.

Digital survey: from new technology to everyday use, a knowledge path and challenge for scholars

Which is the meaning of researching and working in the field of the digital survey at the beginning of the XXIth century? Are the scholars and professionals planning procedures and strategies or just producing an enormous amount of digital data which destiny will be a colossal data loss? Starting from a reflection about “where we are” after 20 years of active digital survey for built heritage this article will try to trace some points about how to start and to plan digital survey intervention when the task is not merely professional and when the new survey bases are supposed to be used in a “liquid” context. From the massive machines and procedures of the XXth century, producing quite “light” amount of data, in the last two decades these tools passed to be lightweight in their hardware, while the amount of gathered data increased continuously, in what it seems an unstoppable process. But massive data gathering maybe it is not knowledge by itself and the information society, especially in its next evolutions, will need contents and versatile data to support and link our present to the heritage values. A specific reflection on the value of digital survey and procedures will be held here not in the pretention of finding a stable paradigm but in the will of stimulating the discussion in a field often tempted by simply technical solutions.

Gujarati Language Model: Word Sense Disambiguation using Supervised Technique

Word Sense Disambiguation (WSD) is a complex problem as it entirely depends on the language convolutions. Gujarati language is a multifaceted language which has so many variations. In this paper, the debate has advanced two methodologies for WSD: knowledge-based and deep learning approach. Accordingly, the Deep learning approach is found to perform even better one of its shortcoming is the essential of colossal data sources without which getting ready is near incomprehensible. On the other hand, uses data sources to pick the implications of words in a particular setting. Provided with that, deep learning approaches appear to be more suitable to manage word sense disambiguation; however, the process will always be challenging given the ambiguity of natural languages

PNNCP- Parallel Nearest Neighbor Classification and Prediction for Big Data Application Based on Apache Spark and Machine Learning

Right by and by the Colossal Information applications, for case, social orchestrating, helpful human administrations, agribusiness, keeping cash, stock show, direction, Facebook and so forward are making the data with especially tall speed. Volume and Speed of the Immense data plays a fundamental bit interior the execution of Colossal data applications. Execution of the Colossal data application can be affected by distinctive parameters. Quickly watch, capacity and precision are the a significant parcel of the triumphant parameters which impact the by and gigantic execution of any Huge data applications. Due the energize and underhanded affiliation of the qualities of 7Vs of Colossal data, each Colossal Information affiliations expect the tall execution.Tall execution is the foremost obvious test within the display advancing condition. In this paper we propose the parallel course of action way to bargain with speedup the explore for closest neighbor center. k-NN classifier is the preeminent basic and comprehensively utilized method for gathering. In this paper we apply a parallelism thought to k-NN for looking the another closest neighbor. This neighbor center will be utilized for putting lost and execution of the remarkable data streams. This classifier unequivocally overhaul and coordinate of the out of date data streams. We are utilizing the Apache Begin and scattered estimation space affiliation for snappier evaluation.

Empowering Internet of Things (IoT) through Big Data

The coming of new advancements, devices and union of remote correspondence, computerized hardware, and miniaturized scale electro-mechanical frameworks (MEMS) advances have brought about the development of Internet of Things (IoT) which thusly delivers an enormous measure of information. IoT helps in diminishing expenses and expanding income, however at the expense of creating tremendous information. The organization of colossal data in an always stretching out framework offers rise to non-irrelevant worries as for data gathering capability, data planning, assessment, and security. This paper describe the benefits of big data in IOT. And also describe architecture and various applications of IOT.