Authors: K. Vigneshwar, P. Shruthi, J. Rahul Naik, P. Khaleel Basha

Abstract: Deoxyribonucleic acid (DNA) is a biological macromolecule. Its main function is information storage. At present, the advancement of sequencing technology had caused DNA sequence data to grow at an explosive rate, which has also pushed the study of DNA sequences in the wave of big data. Moreover, machine learning is a powerful technique for analyzing largescale data and learns spontaneously to gain knowledge. It has been widely used in DNA sequence data analysis and obtained a lot of research achievements. Firstly, the review introduces the development process of sequencing technology, expounds on the concept of DNA sequence data structure and sequence similarity. Then we analyze the basic process of data mining, summary several major machine learning algorithms like Multinomial NB Classifier & Random Forest, and put forward the challenges faced by machine learning algorithms in the mining of biological sequence data and possible solutions in the future. Then we review four typical applications of machine learning in DNA sequence data: DNA sequence alignment, DNA sequence classification, DNA sequence clustering, and DNA pattern mining. We analyze their corresponding biological application background and significance, and systematically summarized the development and potential problems in the field of DNA sequence data mining using Multinomial NB Classifier & Random Forest. Finally, we summarize the content of the review and look into the future of some research directions for the next step.

DOI: https://doi.org/10.5281/zenodo.20425270

Published by: vikaspatanker

Authors: P. Anusha, Syed Mannan Uddin, T.Sree Chandana, V.Vaishnavi

Abstract: Accurate flight delay prediction is fundamental to establish the more efficient airline business. Recent studies have been focused on applying machine learning methods to predict the flight delay. Most of the previous prediction methods are conducted in a single route or airport. This paper explores a broader scope of factors which may potentially influence the flight delay, and compares several machine learning-based models in designed generalized flight delay prediction tasks. To build a dataset for the proposed scheme, automatic dependent surveillance broadcast (ADS-B) messages are received, pre-processed, and integrated with other information such as weather condition, flight schedule, and airport information. The designed prediction tasks contain different classification tasks and a regression task. Experimental results show that long short-term memory (LSTM) is capable of handling the obtained aviation sequence data, but overfitting problem occurs in our limited dataset. Compared with the previous schemes, the proposed random forest-based model can obtain higher prediction accuracy (90.2% for the binary classification) and can overcome the overfitting problem.

DOI: http://doi.org/10.5281/zenodo.20423275

Published by: vikaspatanker

Authors: P. Anusha, E. Naveen Kumar, G. Sravanthi, E. Rohitha

Abstract: Sentiment analysis is a crucial task in natural language processing (NLP) that aims to determine the overall sentiment or opinion expressed by a reviewer towards a movie. This study focuses on the sentiment analysis of IMDB movie reviews using various machine learning and NLP techniques. The findings indicate that feature selection can enhance the accuracy of sentiment-based classification, but the effectiveness depends on the specific method and number of features selected. The paper also presents a comprehensive comparison of traditional machine learning techniques and advanced transformer-based models for sentiment analysis of IMDB movie reviews. The results provide insights into choosing appropriate methods for accurate and timely sentiment analysis on IMDB data. The study employs feature extraction techniques such as bag-of-words (BoW), term frequency-inverse document frequency (TF-IDF), and word2vec. Feature selection using methods like chi-square is shown to improve classification performance.

DOI: http://doi.org/10.5281/zenodo.20423234

Published by: vikaspatanker

Authors: S Jayashree Ananth, Naveen V S

Abstract: The digitization of the healthcare industry has resulted in massive collection of personal health information among hospitals, clinics, and research institutions. But strict privacy laws (HIPAA, GDPR), along with other institutional obstacles, hinder data collection in a centralized manner, resulting in data silos that prevent the construction of efficient machine learning models for predicting diseases, estimating treatment, and managing public health issues. In this paper, we introduce a framework for privacy-preserving federated learning (PPFL) in healthcare. Our proposed framework includes three techniques: (1) Federated Averaging with differential privacy (DP-FedAvg) for model privacy, (2) Secure Multi-Party Computation (SMPC) for private aggregation of gradients, and (3) Homomorphic Encryption (HE) for performing computations on encrypted data. Our PPFL framework is evaluated on three real-life datasets of healthcare applications (mortality prediction from ICU records, diabetic retinopathy classification, and diagnosing COVID-19 patients) and outperforms federated learning with centralization in terms of model accuracy (within 3.2%) and provides differential privacy guarantees with ε=1.0 and δ=10⁻⁵.

DOI: https://doi.org/10.5281/zenodo.20415139

Published by: vikaspatanker

Authors: Dr. Ashish Saxena, Priya Kumari

Abstract: Indias financial world has changed a lot because of technology and digitalization. Now digital payments are a part of our daily transactions. They are fast, easy, convenient and safe. The Unified Payments Interface or UPI is a leading platform for payments. It was developed by the National Payments Corporation of India. UPI helps people transfer money instantly using their phones. They do not need to share their bank details for each transaction. This is possible because many people have smartphones, internet access and use banking. This study looks at how UPI affects users, businesses and Indias digital payment system. It explores how people are changing their behavior to use transactions more. UPI plays a role in helping India become a cashless economy. We look at how UPI simplifies money transfers, bill payments, online shopping and transactions with merchants. We also check how satisfied customers are with UPI, how easy it's to use how fast it is, how secure it is and how reliable it is. The research uses data from questionnaires and surveys well as information from other sources like journals, government reports and websites. We use methods to understand this data and see how UPI affects financial transactions. We found that UPI makes payments more efficient and convenient. It helps reduce the use of cash and increases the use of services. The reasons for this are that UPI is accessible, cost, fast and has many apps. However, there are still some challenges. These include cyber threats, connectivity problems, technical issues and a lack of knowledge. In conclusion UPI has changed Indias payment systems for the better. It has helped include people in the financial system and created opportunities, for businesses, consumers and institutions. As UPI continues to grow it promises to create a secure and cashless financial system that supports a digitally empowered economy.

DOI: http://doi.org/10.5281/zenodo.20412253

Published by: vikaspatanker

Authors: Mr. Shrikant Karampuri, Dr. P. Jeyanthi

Abstract: The fast development of technologies, automation, and digitalization processes have led to a considerable disruption of how workforce planning is traditionally performed and have created a significant gap between the skills of the existing workforce and the skills that will be necessary in the future for the success of organizations. The process of competency mapping which includes the identification, evaluation, and alignment of skills with strategic goals is extremely important to ensure future readiness via skill development. This paper introduces a new approach to the competency mapping using artificial intelligence, which utilizes NLP algorithms for skill extraction from unstructured sources (resumes, job descriptions, and performance evaluations), GNNs for skill adjacency and competency modeling, and BKT for prediction of the evolution of individual skills. When applied to a database of 50,000 employees in a multinational technology company, our approach yields an accuracy of 89.7% for skill extraction, 82% for skill adjacencies, and 76% for the prediction of future skill gaps. The proposed approach allows us to create personalized learning paths and reduce time-to-competency by 34% in six months.

DOI: https://doi.org/10.5281/zenodo.20411976

Published by: vikaspatanker

Authors: Dr. A. Sathiya, Dr. P. Jeyanthi

Abstract: Purchase prediction and understanding its nuances are essential aspects of marketing, but standard approaches do not provide any information about subliminal neural processes which lead to actual purchases. In this paper, a novel multimodal system is proposed which combines EEG neuromarketing data and computer vision features related to visual attention to achieve accurate prediction of consumer purchase intent. A dataset comprising 120 participants who viewed 500 e-commerce images is used for extraction of both EEG-based features (frontal asymmetry of alpha activity, theta/beta ratio, and late positive potential) and visual attention features based on computer vision approach (fixations density, saccades dynamics, and pupils size). Hybrid model consisting of two branches – Temporal Convolutional Network for processing EEG signals and Graph Attention Network for mapping visual attention – reaches 88.3% accuracy and an area under curve equal to 0.94 in predicting consumer purchase intent, while unimodal EEG and visual models reach 74.2% and 72.8% respectively.

DOI: https://doi.org/10.5281/zenodo.20411755

Published by: vikaspatanker

Authors: Dr. Srimathi Kannan

Abstract: The global supply chain network is currently more vulnerable to disruptions that can be caused by pandemics, geopolitics, and other reasons. In such cases, centralized and opaque logistics infrastructures are exposed to risks. This study recommends a reliable supply chain management framework that utilizes blockchain technology, IoT sensors, a hybrid deep learning algorithm to forecast consumer demand, and disruption management through smart contracts. The proposed architecture relies on the Hyperledger Fabric, a permissioned blockchain network, and guarantees data immutability and transparency. A temporal convolutional network with an attention mechanism enables forecasting demand at 95.2% accuracy over a 12-week time frame. After detecting a disruption, the automated smart contract system will engage in dynamic routing, inventory redistribution, and supplier substitution. Simulating the proposed solution on a multi-tier supply chain network with over 100 nodes resulted in 67% faster disruption resolution compared to conventional models and 94% customer satisfaction during disruption events, while conventional models were able to serve just 62% of consumers.

DOI: https://doi.org/10.5281/zenodo.20411670

Published by: vikaspatanker

Authors: Pranit Dattatraya Patil, Avdhoot Arunkumar Sakate, Ajinkya Anil Dhane, Prof. Uchale B. S

Abstract: As cloud-native environments scale, maintaining the high availability of virtual private servers (VPS) has become paramount. Traditional server monitoring tools rely heavily on daemon agents installed on host machines, exposing host environments to resource taxations and security vulnerabilities. This paper presents NexusOps, a premium agentless server management and self-healing platform. By utilizing Java Secure Channel (JSch) tunnels directly to host Operating Systems, NexusOps extracts real-time telemetry metrics (CPU, RAM, Disk, active processes) without target-side exporters. Telemetry metrics are streamed through a centralized Spring Boot REST and WebSocket engine into a responsive React frontend interface, facilitating live command execution, remote service control, and visual analytics. Furthermore, NexusOps introduces a mathemat- ical health-score heuristic model and a multithreaded automated self-healing controller to autonomously resolve critical errors (e.g., service failures, storage spikes) and dispatch alerts via external push notification channels. Our empirical evaluation demonstrates that NexusOps achieves equivalent telemetry ac- curacy and latency (sub-100ms response times) as traditional systems while eliminating persistent CPU and memory footprints on target nodes.

Published by: vikaspatanker

Authors: Vaibhav Mali, Vaibhav Mane, Prajwal Zanje, Prof. N.B.Khade

Abstract: Digital image classification plays a significant role in the early detection and analysis of medical conditions. Traditionally, diagnosis is performed manually by ophthalmologists through examination of retinal fundus images. However, this process is time-consuming, requires expert knowledge, and may sometimes lead to errors due to human limitations. In contrast, automated digital image classification systems provide a faster, more consistent, and cost-effective solution by analyzing large volumes of medical images efficiently. This work focuses on the application of digital image classification techniques for identifying different stages of diabetic retinopathy. Additionally, different image preprocessing, feature extraction, and classification methods are discussed. The study also summarizes the key image features commonly used in previous research for accurate classification of retinal images into different disease categories.

Published by: vikaspatanker