International Journal on Recent and Innovation Trends in Computing and Communication
Not a member yet
8613 research outputs found
Sort by
A Method for Securing Symmetric Keys for Internet of Things Enabled Distributed Data Systems
Get PDFThis study introduces an innovative method for securing symmetric keys in Internet of Things (IoT)-enabled distributed data systems, focusing on enhancing data security while optimizing encryption and decryption times. Through a comprehensive analysis of various encryption algorithms—TEA, XTEA, BLOCK TEA (XXTEA), and the proposed NTSA algorithm—across different key sizes and file sizes, we aim to demonstrate the significant improvements our method offers over existing techniques. Our research meticulously evaluated the performance of these algorithms, employing random variations to encryption and decryption times to simulate real-world variability and assess the algorithms' efficiency and security robustness. The findings reveal that the NTSA algorithm, in particular, showcases superior performance, offering an approximate improvement of 10% to 15% in encryption and decryption times over traditional methods such as TEA and XTEA, and an even more considerable enhancement compared to BLOCK TEA (XXTEA). The key contribution of this study lies in its provision of a secure, efficient framework for symmetric key encryption in IoT-enabled distributed environments. By optimizing key size and algorithm selection, our method not only secures data against potential cyber threats but also ensures high-speed data processing—a critical requirement in the IoT domain where the volume of data transactions and the need for real-time processing are ever-increasing. The proposed method significantly advances the field of data security in distributed systems, especially within the context of the burgeoning IoT landscape. It underscores the importance of algorithmic efficiency and strategic key management in bolstering the security and performance of modern digital ecosystems
An Optimized Genetic Algorithm-Based Non-Commutative Encryption Method for Securing Data in the Cloud
Get PDFThis research introduces a novel non-commutative encryption approach designed to enhance data protection in the context of cloud computing. Leveraging the power of Optimized Genetic Algorithms (OGA), the proposed method aims to fortify the security of sensitive information by introducing non-commutative cryptographic techniques. Cloud computing, while offering unparalleled convenience and scalability, poses inherent security challenges, making robust encryption crucial for safeguarding user data. Through the use of a non-commutative encryption technique, this work presents a novel approach to Quantum Key Distribution (QKD). The integration of genetic algorithms serves to optimize the encryption process, ensuring a balance between computational efficiency and heightened security. There have been several data recovery procedures proposed by researchers, but none of them have shown to be dependable or useful. The suggested method allows users to access data from any backup server if the main cloud server becomes unreliable and cannot provide users with data. In this paper, they perform the analysis based on several parameters such as encryption time, decryption time, success rate, failure rate, throughput, and Avalanche effect. After comparing the proposed work with existing methods, the proposed method has low encryption (312ms)/decryption time (314ms), and a high success rate (100ms)/ failure rate (96ms)
Prediction of Covid-19 Using Fuzzy-Rough Nearest Neighbor Classification
Get PDFPrediction refers to the process of using data and statistical or machine learning techniques to estimate or forecast future events or outcomes based on patterns and trends observed in historical data. The goal of prediction is to make accurate forecasts about what is likely to happen in the future, given what is known about past events and trends. The corona virus has created a global pandemic that significantly disrupted our daily schedule and behaviour patterns. Individuals who contract COVID-19 experience a range of symptoms, which can vary in severity. It is crucial to promptly assess the health condition of individuals affected by COVID-19 by evaluating their symptoms and obtaining essential information. . . To assist in this task, physicians rely on rapid and precise Artificial Intelligence (AI) techniques that aid in predicting patients’ mortality risk and the severity of their conditions. Early identification of a patient’s severity can help conserve hospital resources and prevent patient fatalities by facilitating immediate medical interventions. This research paper introduces an innovative approach that employs the FRNN technique to train a classifier capable of achieving remarkable accuracy in predicting the survival outcomes of COVID-19-affected people. The model is trained on 11 attributes, out of which five are the primary clinical symptoms of this fatal virus: Nasal-Congestion, cough, tiredness, runny nose, fever, sore throat, Diarrhea, and breath shortness, and the other three features are test indication, age, and gender. Our proposed approach, which employs the ENN-SMOTE algorithm to tackle the issue of imbalanced data, demonstrates remarkable effectiveness as evidenced by the experimental results
Cybersecurity Threat Detection Using Machine Learning in Cloud-Based Environments: A Comprehensive Study
Get PDFThe use of cloud computing has made cyb?rs?curity a top priority. Traditional s?curity m?asur?s in dynamic cloud syst?ms rar?ly d?t?ct ?m?rging thr?ats and pr?v?nt th?m from taking action. Th? us? of machin? l?arning algorithms to id?ntify cyb?rs?curity risks in cloud bas?d ?nvironm?nts has b??n ?xplor?d in this ?xt?nsiv? r?vi?w. To configur? risks such as malwar? inf?ctions and p?rsist?nt advanc?d thr?ats and unauthoriz?d acc?ss att?mpts and d?nial of s?rvic? attacks and an int?gration strat?gy that ?.g. mor? vari?ty looks at this sup?rvis?d and unsup?rvis?d and ?ff?ctiv? group l?arning m?thod. Various adv?rsary training t?chniqu?s w?r? us?d to improv? th? r?sili?nc? of th? mod?l to hostil? attacks. This work addresses issues such as data acc?ssibility and mod?l int?rpr?tation and th? dynamic natur? of cyb?r thr?ats and d?monstrat?s th? ?ff?ctiv?n?ss of machin? l?arning in d?t?cting sophisticat?d attacks. It op?ns th? door for s?curity improv?m?nts
Qualities of Learning activities and Illustrations Provisions in the Recommended Chemistry Textbooks for Nigerian Secondary School Students
Get PDFThis study aimed at assessing the qualities of learning activities and illustrations in Nigerian Chemistry textbooks used at the penultimate grades. Five Chemistry textbooks were purposively selected from twenty-two approved and recommended Chemistry textbooks in which evaluation research design was used involving two research questions. The 5 selected Chemistry Textbooks were the instrument for data collection. Data collected were analyzed using QACEST proforma. Results of the study revealed a number of inadequacies in the learning activities and illustrations in these textbooks. It was also observed that none of the Chemistry textbooks emphasized learner centered approach of teaching and the constructivist approach writing
A Review on Identification of Contextual Similar Sentences
Get PDFThe task of identifying contextual similar sentences plays a crucial role in various natural language processing applications such as information retrieval, paraphrase detection, and question answering systems. This paper presents a comprehensive review of the methodologies, techniques, and advancements in the identification of contextual similar sentences. Beginning with an overview of the importance and challenges associated with this task, the paper delves into the various approaches employed, including traditional similarity metrics, deep learning architectures, and transformer-based models. Furthermore, the review explores different datasets and evaluation metrics used to assess the performance of these methods. Additionally, the paper discusses recent trends, emerging research directions, and potential applications in the field. By synthesizing existing literature, this review aims to provide researchers and practitioners with insights into the state-of-the-art techniques and future avenues for advancing the identification of contextual similar sentences
Transforming the Backdrop: Women's Ascension in Indian STEM
Get PDFIn India, women are going through a prominent ascension in the STEM fields, breaking the barricades and are contributing considerably to the nation's technical progression. Their growing existence reflects a transformative change which is apparently authorizing women to prosper in male-dominated domains. The recent times have witnessed a significant progress towards attaining gender equality in the STEM fields. Yet, a there are various challenges that force a women to leave STEM careers at different points during their professional journey. In this study we present an overview of the policies and initiatives for the empowerment of women in STEM in India. We have used the NLP-based content analysis to identify the barriers and challenges rooted in the literature review on women’s empowerment in STEM in India. The research findings help us toformulate the recommendations that determine the explicit research areas and guide our imminent research actions that demand further investigation. The NLP techniques help us to systematically analyze and summarize our literature content to get a better understanding of the women empowerment and thus provide valued insights for the academic and policy-oriented contexts
Blockchain-Driven Logistics Using Ethereum: A Review
Get PDFEveryday life depends heavily on the supply chain, and its traceability guarantees the quality and safety of the products. Thus, there is a pressing need for an effective and trustworthy solution to enhance logistic traceability. Traditional traceability systems suffer from low tracking efficiency and inconsistent data. However, the developing blockchain technology promises to improve these issues by being transparent, tamper-proof, and decentralised. This article analyses previous research, highlights problems, and investigates logistic traceability options based on blockchain. First, the conventional traceability approach and stakeholder demands are explained, along with the fundamentals of blockchain technology. Next, a thorough evaluation and analysis of the current publications and enterprise applications is conducted. Lastly, difficulties and potential lines of inquiry are explored. Subsequent studies may concentrate on developing focused consensus processes, creating suitable access controls, examining the function of regulators in the supply chain, etc. This analysis demonstrates that although there are still many obstacles to overcome, blockchain offers a lot of promise to solve traceability problems
Comparative Study and Framework Design for Twitter Sentiment Detection and Categorization Utilizing Machine Learning Methodologies
Get PDFThe paper is an exhaustive comparative review which gives a way forward in classifying and analyzing sentiments in a twitter data, with the support vector machines (SVM) as the primary analytical tool. The research directs into the depth of sentiment analysis that plays a major role thereby conveying public opinions, trends, and social dynamics of digital platforms like Twitter. Through application of SVN, which is particularly effective in dealing with high dimensional category of data, it gives relatively better knowledge on the context of How Twitter sentiment is tested in the Twitter's short and terse content. The research examines the approaches to sentiment analysis systematically, pointing at vaguely of those methods and of their suitability for the Twitter environment. In particular, this approach underlines that SVM can handle the complexities of Twitter data such as slang words, abbreviations, and emoticons which are quite hard for a text analysis system to manage. The study uses a framework design which is appropriate for the task and takes into account typical operations such as tokenization, stemming, and removal of stop words and this is very crucial for the adjustment of the sensitive model input, using SVM. The results of this empirical assessment will show that SVM is better than NB in some situations, such as when you need to deal with sparse and high-dimensional feature spaces that are typical of Twitter data. The research along with the same walks us through the jeopardy of choosing different kernel functions and finding the ways to set the parameters in SVM which can lead to optimization of the classifier performance. Thus, it also shows an intense path of solving the old challenge. Additionally, the study aims at the demonstration of SVM application in practical scenarios through sentimental analysis. This is to explain how sentiment analysis method can be utilized for business decision-making, political analysis, and social research. It demonstrates the possible effectiveness of SVM-informed sentiment analysis in: assessing the standings of public opinion, diagnosing brand reputation, and the illumination of concerns of the people through the point of view of Twitter.In conclusion, this study not only sheds light on the comparative effectiveness of various sentiment analysis approaches on Twitter but also offers a robust design framework using SVM, contributing valuable insights to the field of text analysis and data mining.
Personalized Health Assessment and Recommendations Through Iot and Mlp Classifier Algorithms
Get PDFProcuring a healthy lifestyle involves a holistic approach of personalized dietary and exercise recommendations dependent on individual health statuses. In this study, we present a new paradigm for examining individual health statuses for easy self-assessment without specialist help. The heart is a full kit of assessing instruments that can align critical climacterics of body temperature, pulse rate, blood oxygen level, and body max index that could be run with minor medic assistance. The research abides a dataset obtained through a broad scope of volunteers aged 17 to 24 including both males and females. Vital signs such as SpO2, BPM, temperature, and BMI are mediated utilizing incorporated Internet of Things units. The dataset is then cautiously preprocessed and balanced using machine learning algorithms before examination. The basis of this model is a two-tier state classifier system that designs autonomous dietary and exercise responsibilities varying from examined health clots. It is exploited for adulthood healthcare systems across multiple machines learning techniques, including Decision Tree, KNN, and some classifiers with the MLP classifier being the exemplary worthy model. The MLP classifier demonstrates unbelievable outcomes through approximately 86% accuracy when the trainings and testing datasets are 70:30 ratios apart