International Journal on Recent and Innovation Trends in Computing and Communication
Not a member yet
8613 research outputs found
Sort by
Nutrition Deficiency Prediction using Machine Learning Techniques
Get PDFDespite the fact that many developing nations have experienced economic progress, Nutrition- deficiency remains a pervasive problem in the society, with millions of impoverished people's diets lacking in essential macro and micronutrients essential for optimal human health. Lack of awareness of food consumed daily causes Nutrition deficiency among general population, data from multiple health records are used for research and prediction. It investigates the importance of a well-balanced diet for our daily life. The Healthy Food Diversity Index (HFDI) is a supplement to the popular Household Dietary Diversity Score (HDDS). It's a tool for determining the diversity of household food. The HDDS has been established as a reliable source of information, but it has several limitations as a measure of dietary diversity that is linked to nutritional quality. In this paper, various machine learning techniques such as Random Forest classifier (RF), Support-Vector Machine (SVM), Linear Discriminant Analysis (LDA) and Logistic Regression (LR) are used to predict Nutrition-Deficiency using house hold risk factors and they compared their Accuracy, Sensitivity and Specificity. The predictions were also compared to the anthropometric classifications used by the National school feeding program to prove the efficiency of the proposed approach
Improvement in Performance of GFDM based 5G Wireless System with Massive MIMO and Channel Coding
Get PDFIn the realm of 5G wireless communication systems, non-orthogonal multiple access (NOMA) waveforms have emerged as highly promising candidates. Various NOMA waveforms such as FBMC, UFMC, and GFDM have gained significant attention due to their numerous advantages when compared to conventional OFDM systems. This paper delves into the examination of Bit Error Rate (BER) performance in the context of a Massive Multiple-Input, Multiple-Output (MIMO) based GFDM system for 5G technology. It explores the effects of different mapping techniques and filter roll-off factors. This approach amalgamates two advanced wireless communication technologies, Massive MIMO and GFDM, with the aim of harnessing their combined potential to elevate the performance and capacity of 5G communication systems. Massive MIMO contributes to substantial improvements in resilience against fading and interference, while GFDM offers superior frequency localization, reduced out-of-band emissions, and enhanced resource allocation flexibility when contrasted with traditional OFDM. The study demonstrates improved results, particularly when employing the optimal roll-off factor for the square cosine filter within the GFDM framework, and integrating channel coding techniques
Diagnosis of Frequency Response Analog Circuits using HHO-SVM
Get PDFMonitoring the system, recognising when a fault has occurred, identifying the kind of defect and where it is located are all aspects of fault detection and isolation. To assess whether a problem has arisen inside a certain channel or region of operation, fault detection is used. For many technological processes in the creation of effective and safe advanced supervision systems, fault detection and diagnosis have grown in significance. This article's main goal is to increase the accuracy of faults detection in frequency response analogue circuits and execution of work needs to be speed up. For this purpose, two optimization techniques are used. One is grey wolf optimization (GWO) for the process of feature extraction and secondly Harris Hawk optimization (HHO) as classifier optimizer. the features and optimize the classifier. The Sallen key circuit (SKC) are utilized for processing the input data. The filters like low pass, high pass and bandpass are designed based on SKC and optimized using GWO. Finally, the optimized features obtained from different circuits are fed to support vector machine classifier to identify the fault accuracy in the circuit. The SVM classifier is optimized using HHO to achieve best accurate output. The suggested technique with a low-dimensional feature optimisation and optimised classifier performed better than the prior methods according to simulation findings, and computing time was also greatly minimised
Automatic Arabic Text Diacritisation Challenges: A Review
Get PDFArabic text diacritization poses a significant challenge within the realm of Natural Language Processing, driven by the intricacies of language modelization and the techniques employed to accurately add diacritics to the text. The complexity is significantly amplified when tackling the Arabic language, characterized by its unique linguistic features. The scarcity of tools and resources dedicated specifi-cally to Arabic diacritization, coupled with the necessity for algorithmic adapta-tion and modelization techniques, further intensifies the challenges. This paper undertakes the task of presenting multiple research endeavors dedicated to Arabic text diacritization, applying diverse algorithms across various datasets. The sub-sequent comparison of these research efforts aims to provide a comprehensive understanding of the field, offering conclusive insights that can guide researchers in their future work and endeavors in this specialized domain
Addressing Persistent Storage Challenges in Kubernetes Environments
Get PDFThe widespread adoption of Kubernetes as a cornerstone for container orchestration highlights the platform's robust capabilities in managing complex, distributed applications. However, an area that consistently presents challenges is the integration of persistent storage solutions. This review paper examines the persistent storage challenges in Kubernetes environments, focusing on the key issues of volume management, data durability, and scalability across diverse infrastructures. Through an extensive literature review, the paper aggregates findings from various studies, revealing common obstacles and innovative solutions in this field. It analyzes how Kubernetes handles stateful applications via Persistent Volumes (PVs), Persistent Volume Claims (PVCs), and StorageClasses, and evaluates the effectiveness of these mechanisms in real-world scenarios. Moreover, the paper delves into the advancements in Container Storage Interface (CSI) drivers and their role in enhancing storage flexibility and operator ease. By synthesizing current research, this review not only outlines the persistent issues but also highlights emerging trends and potential future directions in Kubernetes storage management. It serves as a critical resource for developers, IT professionals, and organizations aiming to leverage Kubernetes for applications requiring reliable and scalable storage solutions
A Novel Fingerprint Recognition and Verification System Using Swish Activation Based Gated Recurrent Unit and Optimal Feature Selection Mechanism
Get PDFUsing fingerprints in biometric systems is a rapidly expanding and pervasive field. The advancement of fingerprint identification as a computer technology for applications is directly linked to the latest developments in computer science. A kind of fingerprint identification algorithm has been made possible by artificial intelligence technology; particularly imaging technology based on deep learning. This paper proposes a novel fingerprint recognition and verification system using a Swish activation-based gated recurrent unit (SWAGRU) with an efficient feature selection mechanism. The system mainly includes four phases: preprocessing, feature extraction, feature selection, and fingerprint recognition. To begin, the fingerprint samples are collected from the publicly available FVC2004 database. After that, Gaussian filtering is applied to the collected dataset to suppress the noise. Then, the feature extraction is carried out with the help of Self-Attention-Based Visual Geometry Group-16 (SAVGG16), and from that, the optimal features are selected based on Cuckoo Search Optimization (CSO). Finally, the fingerprint recognition and verification are done using SWAGRU. The experimental results showed that the system outperformed existing methods in recognition performance
Simulation and Assessment of Bitcoin Prediction Using Machine Learning Methodology
Get PDFThe market for digital currencies is rapidly growing, attracting traders, investors, and businesspeople on a worldwide scale that hasn't been witnessed in this century. By providing comparison studies and insights from the price data of crypto currency marketplaces, it will help in recording the behaviour and habits of such a lucratively demanding and rapidly expanding business. The bitcoin market is reaching one of its peak levels ever in 2021. The emergence of new exchanges has made cryptocurrencies more approachable to the general public, hence boosting their attractiveness. This has increased the number of users and interest in cryptocurrencies, along with a number of reliable crypto ventures started by some of the founders. Virtual currencies are growing more and more well-liked, and businesses like Tesla, Dell, and Microsoft are now embracing them. Decentralized digital currencies are becoming more and more popular, thus it's more crucial than ever to properly inform the public about the new currencies as they proliferate so that people are aware of what they possess and how their money is being invested.
Analysis shows that soft computing and machine learning techniques can anticipate more accurately than any other technique now available to researchers. Finally, it is claimed that ANN, SVMs, and other similar machine learning techniques are useful for predicting global stock market fluctuations.
Analyzing Enterprise Data Protection and Safety Risks in Cloud Computing Using Ensemble Learning
Get PDFNowadays, cloud computing is a significant advancement in the information technology sector. Cloud computing manages and distributes large amounts of data and resources on the internet. In the IT sector, it is used significantly for accessing IT infrastructure via a computer network without necessitating local installations on individual devices. Protecting data security and privacy in cloud computing has become major issue. In this study we used ensemble machine learning algorithms for analysis of cloud computing data, our focus lies analysis of features that effect the data security and privacy threats in cloud computing. Data was gathered through survey online and physical survey method. The data gathering method involved various industries professional’s interactions. The survey dataset features consist of security challenges faced by organizations, such as organization size, industry sector, types of data managed, existing security measures, and prevalent security challenges. The primary focus was on evaluating the effectiveness of three machine learning classifiers: Decision Tree, Random Forest, and Support Vector Machine (SVM), which achieved 85.4%, 89.6%, and 88.2%, accuracies of respectively. To enhance predictive accuracy and robustness, an ensemble learning approach using a voting classifier was implemented, resulting in a significantly improved accuracy of 91.5%. The results show that ensemble learning outperforms individual classifiers in predicting cloud data security threats concerns. This paper highlights significant insights for academics and practitioners by implementing ensemble learning approaches that used for significantly strengthen cloud computing security measures, making them more robust to possible attackers
Secure Cloud Collaboration in Data Centric Security
Get PDFOnline work may be made safer and simpler with the help of secure cloud collaboration. This article discusses challenges encountered, solutions proposed, and novel approaches to data security in cloud collaboration. Access restrictions, data integrity checks, and encryption are some of the instruments that we might employ to secure private data while it is being sent or stored. The report also discusses recent developments that have made cloud collaboration even safer, such as the use of blockchain technology and zero-trust techniques. As more businesses utilize the cloud for collaboration, they must be aware of and abide by security guidelines to preserve client privacy and adhere to legal requirements. This article outlines strategies for enhancing the security of data and cloud collaboration in several businesses, both now and in the future
Integrated Climate Change Impact Analysis
Get PDFThe Integrated Climate Change Impact Analysis project aims to address the pressing issue of pollution-induced health hazards by employing a multifaceted approach. By monitoring and recording the concentrations of various pollutants such as sulphur dioxide, ammonia, carbon monoxide, nitrogen dioxide, nitrogen monoxide, benzene, toluene, xylene, PM10, and PM2.5, the project calculates Air Quality Index (AQI) levels to assess the environmental health risk with respect to time series observations. Mapping these pollutants to associated symptoms and subsequent diseases allows for the prediction of health outcomes, enabling proactive measures to mitigate mortality rates. Through classification techniques such as K-means clustering, the project determines the suitability of cities for habitation based on AQI levels, aiding in pollution reduction strategies. Additionally, machine learning algorithms including Random Forest and Gaussian Naive Bayes are employed to predict diseases from symptoms, facilitating early intervention strategies. Mortality rates are calculated using statistical methods, incorporating probability estimates of disease impact on affected populations. We also aim to categorise if a city is safe to survive by analysing the observed AQI levels. Overall, this project serves as a comprehensive tool to assess environmental health risks, guide urban planning decisions, and ultimately foster healthier living environments