Category Archives: Uncategorized

IJSRET Volume 11 Issue 2, March-April-2025

Uncategorized

Disease Prediction Model Using Multi-Modal Data Fusion
Authors:-Shruti Deokule, Dipti Kause, Suhani Korde, Suhani Korde

Abstract- With recent developments in machine learning and healthcare informatics Strong disease prediction models have been made possible . In order to improve the accuracy and dependability of early disease diagnosis, we present a multi-modal data fusion system in this paper. Advanced fusion techniques that can lessen the drawbacks of single-modality models are used to integrate heterogeneous data sources, such as wearable sensor readings, genomic data, medical images, and electronic health records (EHR). Our method integrates crucial information from multiple datasets by combining feature selection, preprocessing, and ensemble learning. In comparison to the traditional models, we find that the experimental results produce 15% higher prediction accuracy and lower error rates—down to 2.3% for cases of chronic disease.

href=”https://doi.org/10.61137/ijsret.vol.11.issue2.351″>10.61137/ijsret.vol.11.issue2.351

Company Inventory Management System Using Appian
Authors:-Balaji S, Assistant Professor Dr. Krithika. D. R

Abstract- The product in every company decides the availability of resources according to the user needs. Each product must be useful to the user in certain ways to decide as per the demands. This paper speaks about how the products are handled by different departments from storage team to user by choosing the control of each product for the supply and flow in a company. This paper also conveys that this will tell all the activities happens in a single company for deciding how the storage team is very important in storing the products, each team decides the product supply to make it useful for users. When a product gets requested by user it must be decided by the team to inform the availability. The communication mechanism in this application is very useful in understanding the entire system by each team very easily. So, every activity in this application completely named as Inventory to explain about the management of this application.

href=”https://doi.org/10.61137/ijsret.vol.11.issue2.352″>10.61137/ijsret.vol.11.issue2.352

Thermal Analysis of Engine Fins with Different Geometries and Materials
Authors:-Rajat Yadav, Assistant Professor S.N. Dubey

Abstract- In order to make the engine cylinder fins design simpler, CFD (Computational Fluid Dynamics) is used to analyze the thermal and mechanical behavior of the engine cylinder fins. In CFD, the heat transfer and pressure drop characteristics of the engine cylinder fins can be accurately predicted. This helps in understanding the performance of the engine cylinder fins and helps in making necessary modifications to improve the heat dissipation rate. Additionally, CFD can also be used to analyze the airflow characteristics and pressure distribution inside the engine cylinder. This helps in designing the fins properly to reduce the aerodynamic drag and improve the efficiency of the engine. The purpose of this article is to analyze the thermal properties of cylinder fins with different geometries using Ansys Workbench. The geometries were 3D modeled using SOLIDWORKS 2016 and their thermal properties were evaluated using Ansys Workbench R 2016. The change in temperature over time is an important factor in many applications, such as refrigeration, and accurate thermal modeling can help. You determine the key design parameters to improve performance. , The cylinder fin body material is AA 6061 aluminum alloy with a thermal conductivity of 160-170 W/mK. This material is used for current analysis of cylinder ribs.

IoT-Based Electricity Theft Detection System
Authors:-Mohammad Gulrez Zaidi, Deepanshu Punj, Moseen Khan, Ms. Jyoshita Narang

Abstract-Innovative solutions for various industries have been developed as a result of the proliferation of Internet of Things (IoT) devices. IoT has the potential to completely change how electricity is produced, transmitted, and used in the electricity sector. The use of IoT for detecting and preventing electricity theft is one such application. Meter tampering, also known as electricity theft, is a significant problem that affects the revenue and profitability of electricity boards. It entails circumventing meters in an unlawful manner in order to use electricity without paying for it. This not only costs government’s money, but also puts consumers and the electricity grid in danger of injury or damages. In this project, we propose creating an IoT-based system to track down and stop electricity theft. Smart meters with sensors and communication capabilities make up the system, along with a central server for data processing and analysis. Electricity consumption patterns are continuously monitored by smart meters, which also send data to a central cloud-based database. The database values are utilized by the authorities when it discovers anomalies or suspicious activity upon close monitoring of the data stored in real-time. The proposed system could significantly lower the number of instances of electricity theft, increasing revenue and profitability for the electricity providers while enhancing consumer safety. By offering real-time information on electricity consumption and billing, it can also assist utilities in streamlining their operations and enhancing customer service.

DOI: 10.61137/ijsret.vol.11.issue2.357

A Decentralized Social Media Platform with Sentiment Analysis Using Blockchain
Authors:-Nitish Jha, Abhishek Chaudhari, Piyush Pandey

Abstract-This research proposes a decentralized social media platform built on blockchain technology with integrated sentiment analysis using Natural Language Processing (NLP). Traditional social media platforms face issues such as data privacy breaches, central control, and lack of transparency. The proposed system utilizes Ethereum smart contracts and a decentralized architecture to enhance trust, ownership, and user privacy. Sentiment analysis is applied to user-generated content to gain insights and improve user interaction. The system is implemented using the Remix IDE, MetaMask wallet, Sanity database, and ReactJS frontend. This approach provides a transparent, secure, and scalable solution for the next generation of social networking applications.

Welfare Services in Emergency Scenario Management
Authors:-Darshini.S, Assistant Professor Dr. Poongodi.A

Abstract-This Application helps the user to gain the services that are needed for the daily emergency situations. This application provides the services like Petrol, Tow, Hospital, Ambulance and Pharmacy in a Single Module. This application is free for everyone. My goal is to reduce the cost, by providing the maximum service even in any emergency situations to the people needs in a friendly approaching interface. We added braille support so blindly and visually impaired people can also use this application. We added more reliable support system to guide you anytime at anywhere. This application is all in one friendly daily emergency services for people’s welfare.

DOI: 10.61137/ijsret.vol.11.issue2.358

CIFAKE: Image Classification and Explainable Identification of AI-Generated Synthetic Images
Authors:-Assistant Professor Mrs .Shandhini, Roopashree.M, S.Fasiha

Abstract-CIFAKE is a computer program that can detect fake images created by artificial intelligence. It to identifies fake images and provide explanations for its decisions. CIFAKE can help prevent the spread of misinformation and fake news by identifying fake images. This tool can be useful for individuals, organizations, and social media platforms to ensure the authenticity of online images. We introduce the CIFAKE dataset, which consists of 120,000 images (60,000 real images from the CIFAR-10 dataset and 60,000 synthetic images generated using latent diffusion models). Convolutional Neural Networks (CNNs) for binary classification (real vs. fake) and it utilizes Gradient Class Activation Mapping (Grad-CAM) for explainable AI (XAI) to interpret the model’s decisions. The results demonstrate that the given approach achieves a classification accuracy of 92.98%, for detecting AI-generated images.

Solar and Wind Power Electric Vehicle
Authors:-Assistant Professor N. E. K. Chandra., Assistant Professor P.T. Krishna Sai., J. Prudhvi Ganesh.3, Y. Venkata Sai Mohan Krishna., B. Rohit Chandra Sekhar., Md. Saleha.

Abstract-Energy crisis and pollution caused by vehicle emissions are one of the most important issues in the present society. Due to the charging time of battery of electric vehicle, requirement of charging on board is explored as option. This paper deals with the design of a hybrid model of a solar and wind, which uses the battery as its storage system. This system allows the two sources to supply the load separately or simultaneously depending on the availability of the energy sources. The power generated from the wind and solar is fluctuating in nature. The system obtains maximum solar energy during day time and maximum wind energy during the night because the wind blows more at night compared to day time. Therefore, battery of the vehicle can be charged by using hybrid energy system.

A Study on the Challenges in Income Tax Compliance among the Salaried Employees
Authors:-Lavanya S, Assistant Professor Praveen S V

Abstract-This study aims to explore the challenges faced by salaried employees in complying with income tax regulations. Income tax compliance is a significant aspect of the tax system, ensuring the effective functioning of a nation’s economy. However, salaried employees often face various barriers that hinder full adherence to tax laws. These challenges include insufficient knowledge of tax rules, the complexity of tax filing processes, lack of awareness about available exemptions and deductions, and limited access to professional tax advisory services. Additionally, the study examines the impact of digital platforms, such as online filing systems, in facilitating or complicating the compliance process. Data was collected through surveys and interviews with salaried employees from different sectors to identify common issues and concerns. The findings suggest that while digital platforms have made tax filing more accessible, many employees still struggle with understanding tax calculations, deadlines, and documentation. The study also highlights the role of employer assistance in simplifying the compliance process. Recommendations are provided to improve tax literacy, streamline filing procedures, and offer better support systems to enhance overall compliance among salaried employees. This research contributes to understanding the existing barriers in tax compliance and proposes potential solutions to foster greater tax adherence in the salaried workforce.

DOI: 10.61137/ijsret.vol.11.issue2.359

Fitness Club
Authors:-Nirjala Nandekar, Manasi Patil, Vinayak Patil, Chirag Patil, Professor Dr. Prachi Gadhire

Abstract-Being fit physically and mentally is every human being’s ultimate desire. People are always seeking to have a healthy body fitness and they are somehow engaged in day-to-day life. So, we believe that our application can solve this problem in android device users, the apps can be great relief to people who do not have time to visit fitness centre, through help users can manage the healthy life system. Many people who have realized the importance of these apps in their daily life have started making use of such apps. The Current Landscape of the Fitness Apps Market In 2023, the worldwide market for fitness apps clocked in at a solid 1.54 billion USD. Looking ahead, expect this sector to flex its muscles with an impressive compound annual growth rate of 17.7% from 2024 all the way through to 2030. It counts 87.4 million users in the US only. The fitness apps market has evolved dramatically, influenced by changing user preferences and technological advancements. It’s a domain where app users seek personalized experiences, driving the growth of both workout apps and nutrition and diet apps. User Engagement in Mobile Fitness Apps Understanding user engagement in mobile fitness app is crucial. This includes analysing patterns in the Google Play Store and Apple App Store, and how app features cater to diverse user needs. Monetization Strategies: From Paid Apps to In- App Purchases Monetization in the fitness app market varies from paid apps to dynamic in-app purchases, shaping the way developers create revenue streams.

E-Commerce Price Comparison System
Authors:-Assistant Professor Vimmi Malhotra, Kanchan Panwar, Jaya

Abstract-In recent years, mobile apps have become increasingly useful for everyday use. The objective of this project is to provide users with a convenient method to compare product availability and prices across different e-commerce platforms. By inputting the product details into the program, users can effortlessly compare prices from various sources. To compare the product details discovered on multiple websites simultaneously, the application’s databases are searched. To guarantee that they never overlook a fantastic deal, customers can also receive push notifications when items become available or go on sale.

DOI: 10.61137/ijsret.vol.11.issue2.364

AI Chatbot
Authors:-Vipashyna Arun Sable, Professor Suresh Mestry

Abstract-AI chatbots can assist right away by answering questions, providing explanations, and pointing to more resources. Software applications like chatbots can aid teachers in many assignments and become excellent digital teaching assistants.Chatbots are software applications that respond to inputs in natural language. We can see that chatbots are now a part of our daily life. We use them every day to book a movie, reach the closest restaurant, or find an open ATMThese are chatbot software applications that have made life easy. But their uses are not restricted to this. They also entertain users who are bored, play a massive role in home automation projects, give business strategy suggestions, and aid in many other ways.The system was tested on various user inputs and usage scenarios and satisfactory results were observed for usability, accuracy, and responsiveness. The project brings highly evolved chatbot framework technologies like Rasa and chit-chat bots of the likes of facebook and wit.ai to beginner level with a responsive conversational chatbot which could be deployed with minimal coding efforts.The project shows how advanced AI could be used effectively with simple web technologies, and opens up the possibility of further development of using voice, multilingual support, and AI-powered customization.

DOI: 10.61137/ijsret.vol.11.issue2.365

AI Chatbot
Authors:-Vipashyna Arun Sable, Professor Suresh Mestry

Abstract-AI chatbots can assist right away by answering questions, providing explanations, and pointing to more resources. Software applications like chatbots can aid teachers in many assignments and become excellent digital teaching assistants.Chatbots are software applications that respond to inputs in natural language. We can see that chatbots are now a part of our daily life. We use them every day to book a movie, reach the closest restaurant, or find an open ATMThese are chatbot software applications that have made life easy. But their uses are not restricted to this. They also entertain users who are bored, play a massive role in home automation projects, give business strategy suggestions, and aid in many other ways.The system was tested on various user inputs and usage scenarios and satisfactory results were observed for usability, accuracy, and responsiveness. The project brings highly evolved chatbot framework technologies like Rasa and chit-chat bots of the likes of facebook and wit.ai to beginner level with a responsive conversational chatbot which could be deployed with minimal coding efforts.The project shows how advanced AI could be used effectively with simple web technologies, and opens up the possibility of further development of using voice, multilingual support, and AI-powered customization.

Lip-Interpretation Using Deep Learning and Cnn
Authors:-Asmita Chorge, Siddhi Dalvi, Sharvani Mahadik, Ashwini Pawar, Professor Manisha Hatkar

Abstract-Lip interpretation, also known as visual speech recognition, is a challenging task in the field of artificial intelligence (AI) and computer vision. This research explores how deep learning techniques, particularly Convolutional Neural Networks (CNNs), can enhance lip-reading accuracy. By analyzing different architectures, datasets, and methodologies, we present a comprehensive study on various models used for lip interpretation. The findings suggest that deep learning-based approaches significantly improve the accuracy of speech recognition without relying on audio data, thereby opening doors for applications in security, healthcare, and human-computer interaction.

AI Therapist: Artificial Intelligence for Mental Health Support
Authors:- Abinesh M, Prof. Ganapathiram N

Abstract- The emergence of Artificial Intelligence in the field of mental health is transforming traditional approaches to therapy and emotional support. AI Therapist explores the integration of natural language processing, emotion detection, and conversational AI to provide accessible, scalable, and empathetic mental health support. The paper examines current technologies, discusses potential benefits and limitations, and proposes an AI-based mental wellness companion system capable of understanding and responding to emotional states in real-time. As mental health becomes a global concern, AI has the potential to bridge the gap between need and availability while complementing traditional therapy.

Published by:

IoT-Based Smart temperature controlled Fan for Energy-Efficient Cooling

Uncategorized

IoT-Based Smart temperature controlled Fan for Energy-Efficient Cooling
Authors:-Megha Narwade, Diksha Patil, Akanksha Patil, Chetan Aher

Abstract- This paper gives a new and creative IoT-based smart fan system for cooling with energy efficiency in various applications – residential, commercial as well as industrial surroundings. Traditional cooling systems usually depend on static control mechanisms. These cannot adjust to actual environmental conditions and result in extra energy consumption and reduced efficiency too. In order to tackle these challenges, our proposed system integrates an advanced temperature-controlled fan that makes use of a dynamic Pulse Width Modulation (PWM) algorithm that lets us adjust fan speed based on real-time temperature variations around the fan. The system consists of – an NTC thermistor for precise temperature measurement, an Arduino Uno microcontroller for data processing, and an IoT-enabled mobile interface that automates user interactions, and reducing manual adjustments. Strict testing results show that the system successfully achieves a temperature stability of ±0.2°C and reduces power consumption by 27.1% as compared to traditional methods. Also, the modular design supports both DC as well as AC fans, increasing scalability and adaptability. This work bridges the gap between academic research and practical applications and sets a new standard for smart, future focused, thermal management solutions.

DOI: 10.61137/ijsret.vol.11.issue2.340

Published by:

AI-Driven Business Intelligence and Decision Making: Turning Data into Actionable Insights

Uncategorized

AI-Driven Business Intelligence and Decision Making: Turning Data into Actionable Insights
Authors:-Aditya Kokate/strong>

Abstract- The integration of Artificial Intelligence (AI) into Business Intelligence (BI) is revolutionizing how organizations make decisions by automating data analysis and enhancing predictive capabilities. Traditional BI systems rely on static reporting and retrospective data interpretation, limiting their responsiveness in dynamic environments. In contrast, AI-powered BI systems leverage Machine Learning (ML), Natural Language Processing (NLP), and cognitive computing to transform raw data into actionable insights. This paper explores the architecture, implementation, and impact of AI-driven BI systems, emphasizing real-time analytics, predictive modeling, and explainable AI. The study demonstrates how AI enhances operational efficiency, data accuracy, and strategic foresight, offering a competitive advantage to modern enterprises.

DOI: 10.61137/ijsret.vol.11.issue2.339

Published by:

Hand Gesture Recognition in Low-Light Environments Using Deep Learning

Uncategorized

Hand Gesture Recognition in Low-Light Environments Using Deep Learning
Authors:-Sujal Suraywanshi, Ganesh Waje, Diya Jain, Prajyot Jagtap/strong>

Abstract- Hand gesture recognition has attracted enormous interest because of its extensive application in human-computer interaction, sign language understanding, and augmented reality. Recognizing hand gestures in low light conditions is a difficult task to achieve because the visibility is low, there are noises, and feature details get lost. Here, we propose a detailed survey of recent techniques in hand gesture recognition for low-light conditions by employing deep learning methods. We present the difficulties involved with low-light environments, such as illumination changes and background noise. In addition, we discuss different deep learning-based methods for hand detection and gesture classification and present their effectiveness in improving recognition accuracy under difficult lighting conditions. We conclude by offering a comparative assessment of these methods based on primary performance indicators like accuracy, processing time, and resistance to low-light conditions.

DOI: 10.61137/ijsret.vol.11.issue2.338

Published by:

Designing Explainable Large Language Models for Critical Decision-Making in Healthcare: A Review-Based XAI Perspective

Uncategorized

Designing Explainable Large Language Models for Critical Decision-Making in Healthcare: A Review-Based XAI Perspective
Authors:-Roshan Nikam, Parv Shah, Anish Shinde, Chaitanyaa Kashid/strong>

Abstract- Large Language Models (LLMs) are changing the world of healthcare by simplifying clinical documentation, diagnosing, and making medical decisions. Despite their promise, the black-box-like behavior of LLMs poses severe challenges in a healthcare scenario where trust, interpretability, and accountability are of utmost importance. This paper investigates the LLM interpretability techniques in the medical domain based on the salient insights obtained from an extensive survey of Explainable AI (XAI) literature. It looks into post-hoc explanation techniques (SHAP, LIME), collaborative human-AI decision-making frameworks, and interpretable approaches such as neurosymbolic systems. Highlighted are the main issues: algorithmic bias, hallucinations, healthcare compliance, and algorithmic inefficiencies. It is posited that structured prompting, especially Chain-of-Thought (CoT) reasoning enhanced by diagnostic logic, would ensure that LLM actions, in terms of outputs and explanations, are much more in sync with clinical reasoning, thus enhancing transparency while preserving performance. It is concluded that, drawing on the insights gained from XAI, the more interpretable LLMs promote clinicians’ trust in AI systems’ conduct and consequently promote ethical and effective integration into the healthcare setting. The way forward is to focus on some direction to maintain the balance between model accuracy and interpretability and cater to evolving regulatory requirements.

DOI: 10.61137/ijsret.vol.11.issue2.337

Published by:

Free International Journals with High Impact Factor

Uncategorized

International journals are an important source of knowledge, helping to advance research for future generations. Researchers take various steps to publish their work in different repositories, aiming for renowned journals that provide both visibility and recognition. Among them, some authors seek free international journals with high impact factor, which offer the same benefits but at no cost.

Submit Your Paper  / Check Publication Charges

These journals follow strict evaluation measures to ensure that new researchers do not misuse resources and that low-quality papers are removed. High-quality journals that offer free publication often cover their expenses by charging subscription fees from readers rather than from authors.

Journals can be categorized based on free and paid publication charges. However, obtaining all the benefits in both types is not always possible. Paid journals streamline the publication process, making it faster and more efficient. In contrast, free journals often require authors to be patient, as the review process takes longer, and revisions must be made according to reviewer comments.

Some benefits of paid publication include:

  • Responsive publishers with multiple communication channels, including email and phone support.
  • Formatting assistance, allowing authors to focus on research and improving their paper content.
  • Active reviewer follow-ups, with the option to change reviewers if feedback is delayed.
  • Free access to readers, ensuring global visibility, which is a key advantage over free journals.
  • Well-maintained journal portals, updated to meet search engine requirements, resolving technical issues efficiently.

Most mentors serve as reviewers in different journals, but due to their busy schedules, finding time to review papers and follow up on author updates can be challenging. Some journals have a limited pool of reviewers and are constantly in need of new reviewers to verify and evaluate papers. This is one of the limitations of free publication fee journals.

Free Publication Fee Journals

This article helps scholars understand the limitations of free journals and the advantages of paid journals, allowing them to make informed decisions based on their academic and research requirements. Additionally, if a paper is of high quality and covers a novel research topic, it is more likely to be cited by others, further expanding the research domain.

Published by:

Development and Modification of Waste Paper Recycling Machinese

Uncategorized

Development and Modification of Waste Paper Recycling Machinese
Authors:-Mo Mohseen

Abstract- Paper is used daily with learning institutions such as universities and schools being the main consumers. Due to its single usage it ends up being disposed hence most of the paper waste remains idle and unutilized although it is a valuable resource. Therefore, this paper explores the design of a cheap and efficient manually operated paper recycling machine. The design used integration of acquired knowledge on the recycling technology, existing manually operated and available paper recycling machines to form a cheap but efficient paper recycling machine. The benefits of the machine are not only centered on the merits of recycling paper but by the in-cooperation of the manually driving system which will also curb the high unemployment rates in developing countries. Due to the design being not 100% efficient due to the gear box, belt and chain transmission, the estimated efficiency is equal to 90% but using the 90% for design, the design power input is 450 watts and since an average person can produce 100 Watts constantly therefore 2 people are necessary to drive the machine.

DOI: 10.61137/ijsret.vol.11.issue2.336

Published by:

Smart Railway Safety and Track Monitoring System Using ESP32

Uncategorized

Smart Railway Safety and Track Monitoring System Using ESP32
Authors:- Assistant Professor Ambika Annavarapu, Vanja Chaitanya Aswith, Tadiboina Baji, Sattu Naga Likhith, Purushothapatanam Bhanu Sai Ram

Abstract- Railway transportation is one of the most widely used and efficient means of transport worldwide. However, railway accidents due to track failures pose a significant threat to human lives and infrastructure. Cracks and irregularities in railway tracks can cause derailments, leading to catastrophic accidents. To address this problem, we propose a Smart Railway Safety and Track Monitoring System that detects track anomalies using ultrasonic sensors. And ultrasonic sensors are connected to 180° servo motor. The system is designed to continuously monitor the track condition, and if the distance between the track and the ultrasonic sensor exceeds 5 cm, it Indicates a crack or damage. Upon detection of an anomaly, the system triggers a series of safety measures: it sends an alert to the nearest railway station via the Blynk IoT app, transmits a text message with GPS coordinates using a GSM module, and initiates an automated phone call to the concerned authority. Additionally, signal lights positioned beside the track serve as a visual warning, turning from green to red upon detection of a crack. The entire process is controlled by an ESP32 microcontroller, ensuring real-time monitoring and efficient safety measures.

DOI: 10.61137/ijsret.vol.11.issue2.334

Published by:

Engineering Scalable Microservices: A Comparative Study of Serverless Vs. Kubernetes-Based Architectures

Uncategorized

Engineering Scalable Microservices: A Comparative Study of Serverless Vs. Kubernetes-Based Architectures
Authors:- Sreenivasulu Navulipuri

Abstract- This study compares serverless architectures and Kubernetes-based orchestration systems in the context of cloud-native microservices. It examines key factors such as latency, scalability, cost efficiency, and operational complexity. AWS Lambda and other serverless platforms scale automatically but experience delays during cold starts while Kubernetes provides detailed control at the cost of increased operational expenses. The paper also examines AI-driven resource optimization and hybrid models, such as Knative and OpenFaaS, which combine the advantages of both paradigms. Performance benchmarks and case studies guide architects in selecting the most suitable deployment model. A hybrid solution appears to provide the optimal combination of scalability, cost management and operational efficiency.

DOI: 10.61137/ijsret.vol.11.issue2.333

Published by:

Evaluating Machine Learning Efficiency: Simpler Models Outperform Deep Learning in Motor Fault Detection

Uncategorized

Evaluating Machine Learning Efficiency: Simpler Models Outperform Deep Learning in Motor Fault Detection
Authors:-Mrs. L. Yamuna., G. Abhisekhar., K. Prasanna Lahari., K. Vivek., S. Hemanth.

Abstract- In motor condition monitoring, deep learning techniques have been explored by utilizing two-dimensional plots as datasets instead of traditional time-series signals. For instance, Convolutional Neural Networks (CNNs) have been trained using recurrence and frequency-occurrence plots. While previous studies have shown promising results with CNNs, the indistinct differences in these plots often make the model’s decision-making process appear as a black box. This study evaluates and compares ten traditional machine learning (ML) techniques with recent deep learning (DL) approaches for motor fault diagnosis using the same dataset. The dataset consists of 3,750 synthetically generated motor current signal samples, categorized into five classes—one representing healthy conditions and four representing faulty motor conditions—each tested under five loading levels (0%, 25%, 50%, 75%, and 100%). Following similar training and testing phases, the Light Gradient Boosting Machine (LightGBM) achieved the highest classification accuracy of 93.20%, outperforming three CNN-based models by at least 10.4%, whose accuracy ranged between 74.80% and 82.80%. LightGBM also demonstrated superior performance in other key evaluation metrics, including F1 score, precision, and recall. Notably, five out of ten traditional ML models surpassed the CNN-based models. These findings emphasize the importance of carefully selecting deep learning architectures, as they are computationally expensive and memory-intensive, yet do not always guarantee improved performance over traditional ML models, especially for relatively straightforward tasks like motor fault classification using current signals.

DOI: 10.61137/ijsret.vol.11.issue2.328

Published by: