← Back to Archives
VOLUME 14, ISSUE 9, SEPTEMBER 2026
Publishing in progress..
Valorization of Unripe Plantain (Musa paradisiaca) Peel for a Renewable Bioethanol Production
Abdulsalam Abdulrahman Ayodeji, Gbadamosi L, Adejumo Mutiu, Saliyu M, Haliru M, Dan illela U. Y
IOT Based Hand-Gesture Controlled LED Blinking Effect
Siddharth S. Birajdar, Vaibhav K. Deshmukh, Gayatri P. Chandane, S. T. Sanamdikar
Simulation-Based Investigation of Locally Available Natural Materials as Dielectric Substrates for Microstrip Patch Antennas
Jamilu Mohammed Zuru, Shamsudeen Umar, Saidu Sunbo Akanji
Energy-Efficient CMOS Ring Oscillator for Sustainable 5G Communication System: A Comprehensive Literature Review
Basavaraju S Hugar, Rakshitha R, Chethan K N, Prajwal Swamy H S, Dr.N.C.Patil
Biomechanical Analysis of the Sprint-Chase (Pursuit Running) Skill in Kho-Kho
Dr. Pathan Yusufkhan Hidayatkhan
CYBER SECURITY AND ITS GUIDELINES
Mr. Tushar S. Magar, Prof. Miss. Chetana M. Kawale
An AI-Driven Framework for Inclusive and Personalized Skill Assessment
Mr. Gaurav Kumbhar, Prof. Chetana M Kawale
IMPACT ASSESSMENT OF ROTIFER AND COPPENS ON THE DEVELOPMENT OF AFRICAN CATFISH (Clarias gariepinus) LARVAE
Umar Abubakar*, Abdulrahman Dantani, Oladele Sunday Mathew, Alhassan Abdullahi
A Scalable Deep Learning Framework for Diabetic Retinopathy Classification with Web Integration and FPGA-Based Edge Deployment
Ram Rishik Nalukurthi, Ramadevi Kolisetty
DIGITAL TECHNOLOGY TO CALCULATE WATER FOOTPRINTS FOR DIFFERENT AGRICULTURAL PRODUCTS
Mr. Bhagyesh Shamal Fegade, Asst.Prof.Ms.Sapana A.Fegade
DATA DOWNLOAD DUPLICATION ALERT SYSTEM
Miss. Pratiksha S. Hiwrale, Dr. Sarkarsinha H. Rajput, Dr. Shital A. Patil, Asst. Prof. Arsalan A. Shaikh
An Integrated Approach for 2D Blueprint Understanding and 3D Model Reconstruction
Miss. Pallavi R. Gaikwad, Dr. Sarkarsinha H. Rajput, Dr. Shital A. Patil, Asst. Prof. Arsalan A. Shaikh
AI-DRIVEN CROP DISEASE PREDICTION AND MANAGEMENT SYSTEM
Miss.Radhika Sugdev Gavhale, Asst. Prof. Arsalan A. Shaikh
Monitoring System for Classroom Session in Skill Training
Miss. Vaishnavi D. Ghadekar, Asst. Prof. Mr. Arsalan A. Shaikh
HEALTH DATA INFORMATION & MANAGEMENT SYSTEM
Miss. Riya D. Kachare, Asst. Prof. Arsalan A. Shaikh
Fake Social Media Accounts and Their Detection
Miss. Gayatri R. Gatmane, Asst. Prof. Arsalan A. Shaikh
AI-OPTIMIZED TIME-SLOT DELIVERY FOR EFFICIENT PARCEL DISTRIBUTION IN E- COMMERCE LOGISTICS
Miss. Sanika Bhika Jaware, Asst. Prof. Arsalan A. Shaikh
DEVELOP A DASHBOARD AIMED AT MAINTAIN CLEAN AND GREEN TECHNOLOGY
Miss. Vidya R. Ingole, Asst. Prof. Arsalan A. Shaikh
Vehicle-to-Grid (V2G) Enabled Renewable-Powered Microgrids: A Sustainable Approach to Carbon Emission Reduction
Suraj Chaudhary, Himanshu Tomar, Jasmine Kaur
Design and Performance Analysis of Battery–Supercapacitor Hybrid Energy Storage System with PSO-Based PI Controller Optimization
Himanshu Tomar, Suraj Chaudhary, Jasmine Kaur
Development AI/ML Based Models for Predictions Prices for Agri-Horicultural Commodities Such as Pulses and Vegetables (Onion, Potato, Onion)
Miss. Dhanshree G. Mahajan, Asst. Prof. Miss. Chetana M. Kawale
Identification Of Algorithm from the Given Dataset Using AI/ML Techniques
Miss. Sakshi D. Mahajan, Asst. Prof. Chetana M. Kawale
Enhancing Rail Madad AI-Powered Complaint Management System
Mr. Kalpesh K. Manure, Asst. Prof. Chetana M. Kawale
Reporting Uncertainty in Permutation Feature Importance: Why the Usual Dispersion Measure Is the Wrong One
Ofierohor Ufuoma Earnest
Abstract
Valorization of Unripe Plantain (Musa paradisiaca) Peel for a Renewable Bioethanol Production
Abdulsalam Abdulrahman Ayodeji, Gbadamosi L, Adejumo Mutiu, Saliyu M, Haliru M, Dan illela U. Y
DOI: 10.17148/IJIREEICE.2026.14901
Abstract: The increasing demand for sustainable alternatives to fossil fuels has stimulated interest in lignocellulosic biomass as a renewable feedstock for bioethanol production. This study assessed the potential of unripe plantain (Musa paradisiaca) peels as a lignocellulosic substrate for bioethanol production, with emphasis on physicochemical composition, reducing sugar recovery, inhibitor reduction and ethanol yield. Fresh unripe plantain peels were washed, sundried, milled and subjected to combined pretreatment using a combination of 10% formic acid and 5% tetraoxosulphate vi acid at 121oC for 30min, followed by acid hydrolysis. The physicochemical composition of the raw and pretreated biomass were determine using standard methods. The raw unripe plantain peel contained 54.34% carbohydrate, 28.60 cellulose, 17.40% hemicelluloses and 13.80% lignin. Following acid pretreatment, cellulose increased to 48.63.%, while hemicelluloses and lignin decreased to 8.63% and 7.46% respectively, indicating effective disruption of the biomass structure. The acid hydrolysis of the pretreated biomass produced a hydrolysate with 21.59g/l reducing sugars, corresponding to 51.80% sugar recovery. Detoxification of the hydrolysate using overliming and activated at 1:3 and 3:1 ratios reduces the inhibitor index by 40,98% and 54.76% respectively, with 3:1 ratio showing greater detoxification efficiency. Fermentation of the hydrolysate using yeast produced an ethanol concentration of 3.82g and 14.86% yield for undetoxified sample, 6.43g and 38.91% yield at 1:3 ratio, and 8.42g and 50.18% at 3:1 detoxification ratio. The findings demonstrate that unripe plantain peels possesses considerable potential as a renewable lignocellulosic feedstock for bioethanol production. Valorization of unripe plantain peel waste contributes to waste management, renewable energy generation, and sustainable biofuel development.
Keywords: Lignocellulosic biomass, Renewable energy, unripe plantain peels, Fossil fuel, Bioethanol
Keywords: Lignocellulosic biomass, Renewable energy, unripe plantain peels, Fossil fuel, Bioethanol
Abstract
IOT Based Hand-Gesture Controlled LED Blinking Effect
Siddharth S. Birajdar, Vaibhav K. Deshmukh, Gayatri P. Chandane, S. T. Sanamdikar
DOI: 10.17148/IJIREEICE.2026.14902
Abstract: The fast growth of human-machine interaction has created a need for simple, easy-to-use, and contactless ways to control electronic devices.
Traditional control systems usually rely on physical switches, buttons, or wired connections, which might not always be convenient or flexible. This research introduces a low-cost Hand-Gesture Controlled LED Blinking Effect system that lets users control LED lighting and blinking patterns using specific hand gestures. The system uses an Arduino Uno microcontroller, HC-05 Bluetooth module, LED lights, resistors, and a breadboard to offer wireless control of the LED circuit. In this approach, hand gestures are recognized as set control commands and sent wirelessly via the HC-05 Bluetooth module.
The Arduino Uno receives and processes these commands based on the programmed logic. Depending on the received command, the microcontroller controls the connected LEDs and creates the appropriate blinking effect. Various commands can be linked to different LED actions, such as turning the LED on or off, adjusting the blinking speed, or forming blinking patterns. This method removes the need for direct physical contact with the LED control setup and provides an easy way to show gesture-based device control.
The system emphasizes simplicity, affordability, ease of use, and dependable wireless communication. The experimental setup shows that an Arduino-based platform can effectively handle Bluetooth-based gesture commands and control LED outputs with minimal hardware. The proposed system serves as a fundamental model for understanding gesture-based human-machine interaction, embedded systems, and wireless control technologies. Also, the idea can be expanded to control multiple LEDs and other electronic devices by adding more control commands and appropriate hardware. Possible uses include smart lighting systems, home automation, assistive technology, interactive electronic devices, educational projects, and touchless control systems. This work highlights the potential of combining gesture- based interaction with affordable microcontrollers and Bluetooth technology to build accessible and scalable electronic control systems.
Keywords: Hand Gesture Recognition, Arduino Uno, HC-05 Bluetooth Module, LED Control, Wireless
Traditional control systems usually rely on physical switches, buttons, or wired connections, which might not always be convenient or flexible. This research introduces a low-cost Hand-Gesture Controlled LED Blinking Effect system that lets users control LED lighting and blinking patterns using specific hand gestures. The system uses an Arduino Uno microcontroller, HC-05 Bluetooth module, LED lights, resistors, and a breadboard to offer wireless control of the LED circuit. In this approach, hand gestures are recognized as set control commands and sent wirelessly via the HC-05 Bluetooth module.
The Arduino Uno receives and processes these commands based on the programmed logic. Depending on the received command, the microcontroller controls the connected LEDs and creates the appropriate blinking effect. Various commands can be linked to different LED actions, such as turning the LED on or off, adjusting the blinking speed, or forming blinking patterns. This method removes the need for direct physical contact with the LED control setup and provides an easy way to show gesture-based device control.
The system emphasizes simplicity, affordability, ease of use, and dependable wireless communication. The experimental setup shows that an Arduino-based platform can effectively handle Bluetooth-based gesture commands and control LED outputs with minimal hardware. The proposed system serves as a fundamental model for understanding gesture-based human-machine interaction, embedded systems, and wireless control technologies. Also, the idea can be expanded to control multiple LEDs and other electronic devices by adding more control commands and appropriate hardware. Possible uses include smart lighting systems, home automation, assistive technology, interactive electronic devices, educational projects, and touchless control systems. This work highlights the potential of combining gesture- based interaction with affordable microcontrollers and Bluetooth technology to build accessible and scalable electronic control systems.
Keywords: Hand Gesture Recognition, Arduino Uno, HC-05 Bluetooth Module, LED Control, Wireless
Abstract
Simulation-Based Investigation of Locally Available Natural Materials as Dielectric Substrates for Microstrip Patch Antennas
Jamilu Mohammed Zuru, Shamsudeen Umar, Saidu Sunbo Akanji
DOI: 10.17148/IJIREEICE.2026.14903
Abstract: This research presents a simulation-based performance investigation of a 5.8 GHz microstrip patch antenna using locally available natural materials as dielectric substrates. Natural substrates, including Bamboo, Wood, Paper, Cotton and Palm fiber, were characterized alongside standard FR-4 dielectric material. The analytical framework evaluated fundamental dielectric parameters, including relative permittivity (εr), loss tangent (tan δ), imaginary permittivity (ε″) and effective permittivity (εeff), revealing an inverse relationship between εr and required patch footprint (ranging from 15.798 mm × 11.803 mm for FR-4 up to 22.683 mm × 19.139 mm for Cotton). Closed-form transmission line equations produced an initial analytical frequency offset of 9.58% to 12.28% above the 5.8 GHz target, necessitating numerical co-optimization in CST Studio Suite. Benchmark validation of custom FR-4 (εr = 4.36, tan δ = 0.01) against CST’s built-in library model confirmed numerical model fidelity across 4–7 GHz. Full-wave CST simulations demonstrated that Cotton (εr = 1.6, tan δ = 0.0015) achieved superior performance with the highest radiation efficiency (67.0%) and peak gain (3.66 dBi), outperforming all other natural dielectrics. Wood and Palm fiber provided strong impedance matching (S11 = −35.31 dB and −33.36 dB; VSWR = 1.35 and 1.13) with moderate efficiencies (45.8% and 41.2%). Paper exhibited the highest attenuation (tan δ = 0.04), yielding the lowest radiation efficiency (29.8%) and peak gain (1.45 dBi). These findings establish the viability of natural dielectrics for eco-friendly wireless systems while delineating key physical trade-offs in substrate loss and antenna geometry.
Keywords: Microstrip patch antenna, Natural dielectric substrate, 5.8 GHz ISM band,CST Studio Suite,Radiation efficiency, Eco-friendly antenna
Keywords: Microstrip patch antenna, Natural dielectric substrate, 5.8 GHz ISM band,CST Studio Suite,Radiation efficiency, Eco-friendly antenna
Abstract
Energy-Efficient CMOS Ring Oscillator for Sustainable 5G Communication System: A Comprehensive Literature Review
Basavaraju S Hugar, Rakshitha R, Chethan K N, Prajwal Swamy H S, Dr.N.C.Patil
DOI: 10.17148/IJIREEICE.2026.14904
Abstract: The rapid deployment of fifth-generation (5G) wireless communication systems has significantly increased the demand for compact, high-frequency, low-power and energy-efficient radio-frequency integrated circuits. Frequency generation is one of the most critical functions in a 5G transceiver because phase-locked loops (PLLs), frequency synthesizers, clock generators and local oscillators directly influence system power consumption, phase noise, frequency stability and communication performance. Conventional LC-tank voltage-controlled oscillators (LC-VCOs) provide excellent phase-noise performance but suffer from large silicon area, limited frequency tuning flexibility and relatively high implementation complexity. CMOS ring oscillators and ring-based voltage-controlled oscillators (RO-VCOs), in contrast, provide attractive advantages in terms of compact area, wide tuning range, digital compatibility and low power consumption.
Recent IEEE research has demonstrated substantial progress in ring-oscillator-based frequency generation for wireless
jitter limitations of ring oscillators remain important challenges for 5G applications. Recent approaches include current- starved delay cells, pseudo-differential architectures, digitally controlled ring oscillators, injection locking, multi-phase generation, adaptive tuning, supply-noise suppression and advanced CMOS/FinFET technologies. This paper systematically reviews these developments and analyzes the trade-offs among power consumption, oscillation frequency, tuning range, phase noise, jitter, area and figure of merit (FoM).
Particular attention is given to the relationship between oscillator energy efficiency and sustainable 5G communication. A comparative analysis of recent IEEE publications is presented, followed by a taxonomy of CMOS ring-oscillator architectures and identification of important research gaps. The review indicates that future energy-efficient 5G/6G ring oscillators should combine adaptive power management, low-voltage operation, digitally assisted calibration, improved phase-noise suppression and technology-aware delay-cell optimization.
Keywords: CMOS, ring oscillator, ring VCO, voltage-controlled oscillator, low power, energy efficiency, 5G, 6G, PLL, phase noise, jitter, frequency synthesizer, sustainable communication, VLSI.
Recent IEEE research has demonstrated substantial progress in ring-oscillator-based frequency generation for wireless
jitter limitations of ring oscillators remain important challenges for 5G applications. Recent approaches include current- starved delay cells, pseudo-differential architectures, digitally controlled ring oscillators, injection locking, multi-phase generation, adaptive tuning, supply-noise suppression and advanced CMOS/FinFET technologies. This paper systematically reviews these developments and analyzes the trade-offs among power consumption, oscillation frequency, tuning range, phase noise, jitter, area and figure of merit (FoM).
Particular attention is given to the relationship between oscillator energy efficiency and sustainable 5G communication. A comparative analysis of recent IEEE publications is presented, followed by a taxonomy of CMOS ring-oscillator architectures and identification of important research gaps. The review indicates that future energy-efficient 5G/6G ring oscillators should combine adaptive power management, low-voltage operation, digitally assisted calibration, improved phase-noise suppression and technology-aware delay-cell optimization.
Keywords: CMOS, ring oscillator, ring VCO, voltage-controlled oscillator, low power, energy efficiency, 5G, 6G, PLL, phase noise, jitter, frequency synthesizer, sustainable communication, VLSI.
Abstract
Biomechanical Analysis of the Sprint-Chase (Pursuit Running) Skill in Kho-Kho
Dr. Pathan Yusufkhan Hidayatkhan
DOI: 10.17148/IJIREEICE.2026.14905
Abstract: Kho-Kho is a traditional Indian tag-based team sport in which a chaser's ability to pursue and tap an opposing runner decides the outcome of a turn. The central skill of this pursuit -- repeated short-burst sprinting, rapid acceleration from a crouched “kho”-receiving posture, and sustained high-velocity chase across the playing lane -- has rarely been examined from a mechanical standpoint despite its decisive tactical role. This paper analyses the sprint-chase skill of kho-kho through the lens of linear kinetics and kinematics, treating the chaser's body as a system governed by Newton's laws of motion, the impulse-momentum relationship, and stride mechanics. Using a combined framework of high-speed two-dimensional video analysis and force-platform-derived ground reaction force (GRF) data, the study models the phases of the chase -- reaction and start, acceleration, and near-maximal pursuit -- and identifies the propulsive and braking force components that govern stride length, stride frequency, and horizontal acceleration of the centre of mass. Fifteen state-level male kho-kho chasers (18-24 years) were filmed executing the sprint-chase skill under match- simulated conditions, and stride kinematics were digitised frame by frame and compared between successful and unsuccessful pursuit trials. Results indicate that effective chasers generate a significantly greater horizontal propulsive impulse during the first three strides after the kho is received, shorter ground contact time during acceleration, and a forward trunk lean of approximately 45° at push-off that optimises the horizontal component of GRF. Stride frequency, more than stride length, was the stronger correlate of a completed tap. These findings position the sprint-chase as a mechanically distinct skill shaped by the game's repeated stop-start structure, offering a biomechanical basis for chaser- specific conditioning that favours acceleration mechanics over open top-end speed.
Abstract
CYBER SECURITY AND ITS GUIDELINES
Mr. Tushar S. Magar, Prof. Miss. Chetana M. Kawale
DOI: 10.17148/IJIREEICE.2026.14906
Abstract: Cyber security has become one of the most critical concerns in the modern digital era, as organizations, governments, and individuals are increasingly reliant on information technology for daily operations and communication. The rapid proliferation of internet-connected devices and cloud-based services has expanded the threat landscape significantly, exposing sensitive data and infrastructure to a wide range of cyber threats. This paper presents a comprehensive study on cyber security and its guidelines, focusing on the various types of cyber attacks including phishing, malware, ransomware, Distributed Denial of Service (DDoS), Man-in-the-Middle (MitM), SQL injection, and social engineering attacks. The study examines existing security frameworks and guidelines such as the National Institute of Standards and Technology (NIST) Cyber Security Framework, ISO/IEC 27001, and the Cyber Essentials scheme. A systematic literature review methodology has been adopted to analyze recent research on cyber attack patterns, vulnerability assessment techniques, and defence mechanisms. The results reveal that human factors remain the weakest link in cyber security, with phishing accounting for over 36% of all data breaches. The paper proposes a set of practical guidelines and best practices for individuals and organizations to protect against prevalent cyber threats. The findings suggest that a combination of technical controls, user awareness training, and adherence to established security frameworks can significantly reduce the risk of cyber attacks.
Keywords: Cyber Security, Cyber Attacks, Phishing, Malware, Ransomware, NIST Framework, Security Guidelines, Data Protection
Keywords: Cyber Security, Cyber Attacks, Phishing, Malware, Ransomware, NIST Framework, Security Guidelines, Data Protection
Abstract
An AI-Driven Framework for Inclusive and Personalized Skill Assessment
Mr. Gaurav Kumbhar, Prof. Chetana M Kawale
DOI: 10.17148/IJIREEICE.2026.14907
Abstract: Assessment is a critical component of education, vocational training, and workforce skill development, yet conventional assessment systems continue to depend on fixed question sets, uniform response formats, and manual evaluation. These practices do not adequately account for differences in learner ability, preferred assessment mode, or accessibility requirements. this paper proposes and defines an evaluable AI-driven inclusive assessment framework that couples adaptive question selection, automated evaluation of objective and descriptive responses, accessibility support, skill-gap analysis, and performance analytics within a single platform, while retaining a human reviewer for uncertain or high-impact decisions. The framework supports multiple-choice, descriptive, practical, and viva-based assessment modes for candidates with a range of abilities and accessibility needs.The methodology follows nine interconnected stages, from requirement analysis and candidate modelling through adaptive selection, automated evaluation, accessibility support, skill-gap analysis, analytics, human review, and privacy governance.against which the enhanced pipeline can be compared, and evaluation is organised around measurable, component-wise criteria: question-selection efficiency and skill-estimation consistency for adaptive testing; scoring agreement (e.g., Cohen's kappa) and error analysis for automated evaluation; usability and accessibility-issue counts for the accessibility layer; and response-time and throughput for system performance.; instead, the exact baseline, metrics, and comparison procedure that a subsequent implementation will use are defined precisely. The principal contribution is not any single AI technique but the integration of previously separate capabilities into one accountable, measurable, and inclusive assessment workflow suitable for a diverse skill ecosystem.
Keywords: AI-Driven Assessment, Inclusive Assessment, Adaptive Testing, Automated Evaluation, NLP, Skill-Gap Analysis, Accessibility, Human-in-the-Loop
Keywords: AI-Driven Assessment, Inclusive Assessment, Adaptive Testing, Automated Evaluation, NLP, Skill-Gap Analysis, Accessibility, Human-in-the-Loop
Abstract
IMPACT ASSESSMENT OF ROTIFER AND COPPENS ON THE DEVELOPMENT OF AFRICAN CATFISH (Clarias gariepinus) LARVAE
Umar Abubakar*, Abdulrahman Dantani, Oladele Sunday Mathew, Alhassan Abdullahi
DOI: 10.17148/IJIREEICE.2026.14908
Abstract: The African catfish (Clarias gariepinus) is the most widely cultivated fish species in Nigeria due to its fast growth rate, high fecundity, disease resistance, and adaptability to artificial feed. However, Nigerian fish farmers face significant challenges, particularly the high cost and limited availability of quality feed, which accounts for approximately 60% of production costs. This study compared the suitability of live rotifers (Brachionus spp.) and Coppens, a commercial fish feed, for feeding C. gariepinus larvae. The experiment was conducted at the Fisheries Hatchery Unit, Kebbi State University of Science and Technology, using a completely randomized design (CRD) with six replicates. Growth performance indicators such as weight gain, length gain, specific growth rate (SGR), feed conversion ratio (FCR), and survival rate were evaluated. The results showed that C. gariepinus fed Coppens exhibited significantly higher growth performance (P<0.05) than those fed live rotifers, with superior weight gain (0.80±0.28 g vs. 0.66±0.26 g), length gain (3.04±0.34 cm vs. 2.76±0.33 cm), and SGR (2.91±0.16% vs. 1.28±0.90%). Conversely, fish fed rotifers showed a higher survival rate (68.70±3.69%) compared to those fed Coppens (53.00±7.00%). The findings indicate that while Coppens is more effective for growth, live rotifers can serve as a supplementary feed. It is recommended that rotifers can be further processed into rotifer meal to enhance their nutritional value and reduce dependence on expensive imported fishmeal.
Keywords: AFRICAN CATFISH, Clarias gariepinus, ROTIFERS, COPPENS, LARVAE.
Keywords: AFRICAN CATFISH, Clarias gariepinus, ROTIFERS, COPPENS, LARVAE.
Abstract
A Scalable Deep Learning Framework for Diabetic Retinopathy Classification with Web Integration and FPGA-Based Edge Deployment
Ram Rishik Nalukurthi, Ramadevi Kolisetty
DOI: 10.17148/IJIREEICE.2026.14909
Abstract: Diabetic retinopathy (DR) is a progressive microvascular complication of diabetes and a major cause of preventable vision loss. Early and accurate identification of DR severity is essential; however, conventional screening relies heavily on manual examination of retinal fundus images, making large-scale diagnosis time-consuming and dependent on specialist availability. This work presents an end-to-end framework for automated DR classification that combines deep learning, web-based accessibility, and FPGA-oriented hardware implementation. A customized residual convolutional neural network is developed to classify retinal fundus images into five severity levels: No DR, Mild, Moderate, Severe, and Proliferative DR. The model employs residual skip connections to improve feature propagation and support effective learning of pathological retinal characteristics. A preprocessing and augmentation pipeline is incorporated to standardize image representation, improve feature visibility, and address class imbalance. The proposed model achieves an overall classification accuracy of 97.2%, demonstrating strong performance in multi-class DR classification. For practical accessibility, the trained model is integrated into a Flask-based web application that enables users to upload retinal images and obtain automated predictions. To extend the framework toward edge-oriented deployment, the convolutional parameters are converted from floating-point to fixed-point integer representation and mapped to a custom FPGA accelerator on the Zynq-7000 XC7Z010-1CLG400 platform. The hardware architecture incorporates parallel 3×3 multiply–accumulate processing, bias addition, re-quantization, and ReLU activation, with parameter and data movement supported through the Zynq processing system, DDR memory, and AXI/DMA interfaces. The proposed framework therefore establishes a unified path from software-based DR classification to web-accessible screening and quantized FPGA implementation, providing a foundation for efficient edge-based medical image inference.
Keywords: Diabetic Retinopathy, Deep Learning, Residual Convolutional Neural Network, Fundus Image Classification, Multi-Class Classification, Quantization, FPGA, Zynq-7000, Edge AI, CNN Accelerator, Web-Based Screening.
Keywords: Diabetic Retinopathy, Deep Learning, Residual Convolutional Neural Network, Fundus Image Classification, Multi-Class Classification, Quantization, FPGA, Zynq-7000, Edge AI, CNN Accelerator, Web-Based Screening.
Abstract
DIGITAL TECHNOLOGY TO CALCULATE WATER FOOTPRINTS FOR DIFFERENT AGRICULTURAL PRODUCTS
Mr. Bhagyesh Shamal Fegade, Asst.Prof.Ms.Sapana A.Fegade
DOI: 10.17148/IJIREEICE.2026.14910
Abstract: Agricultural production is inextricably linked with freshwater consumption, rendering quantitative, reliable water-footprint assessments essential for sustainable environmental resource management. This research proposes a robust digital technology-based framework for systematically calculating, validating, organizing, and visualizing the water footprints of diverse agricultural commodities. The proposed digital workflow integrates field data collection, structured product-level digital records, component-wise water-footprint computation engines, automated validation rules, and interactive reporting dashboards. Green (rainwater), blue (surface and groundwater), and grey (pollutant dilution volume) water footprint components are quantified separately and combined to yield precise product-level water footprint indicators (m³/ton). Digital record management drastically improves data traceability by preserving raw data sources, mathematical calculation methods, model version parameters, timestamps, and data quality flags. Operating within a design-oriented engineering framework, this study defines the architectural layers, database schema attributes, sequential data workflows, and a target performance evaluation matrix. The resulting framework provides a transparent foundation for multi-crop comparison, policy reporting, and future integration with Internet of Things (IoT) field sensors, remote sensing, artificial intelligence, and smart agriculture decision-support ecosystems.
Keywords: Water footprint assessment, Digital agriculture, Green-blue-grey water, Data traceability, Precision farming, Smart agriculture, Quality validation, IoT integration.
Keywords: Water footprint assessment, Digital agriculture, Green-blue-grey water, Data traceability, Precision farming, Smart agriculture, Quality validation, IoT integration.
Abstract
DATA DOWNLOAD DUPLICATION ALERT SYSTEM
Miss. Pratiksha S. Hiwrale, Dr. Sarkarsinha H. Rajput, Dr. Shital A. Patil, Asst. Prof. Arsalan A. Shaikh
DOI: 10.17148/IJIREEICE.2026.14911
Abstract: The Data Download Duplication Alert System is a software system developed to help users avoid downloading the same file multiple times. Today, people frequently download documents, images, videos, software, and other files from websites, cloud platforms, and online repositories. Sometimes, the same file is downloaded again without the user realizing it. This can take up extra storage space, use more internet bandwidth, and waste time.The main purpose of this research is to develop a system that can check whether a file has already been downloaded and notify the user before downloading it again. The system keeps track of previously downloaded files and checks information such as the file name, size, type, and hash value. If the requested file is found to be a duplicate, the system displays an alert to the user. The user can then choose whether to continue with the download or cancel it.The system can be developed using Python, databases, and file hashing techniques. By detecting duplicate files, the system can help reduce unnecessary downloads, save storage space, and decrease network bandwidth usage. It can also make file management easier for users.The proposed system can be useful for individual users, educational institutions, and organizations. In the future, it can be improved by adding cloud storage integration, smarter duplicate detection, real-time monitoring, and automatic file management features.
Abstract
An Integrated Approach for 2D Blueprint Understanding and 3D Model Reconstruction
Miss. Pallavi R. Gaikwad, Dr. Sarkarsinha H. Rajput, Dr. Shital A. Patil, Asst. Prof. Arsalan A. Shaikh
DOI: 10.17148/IJIREEICE.2026.14912
Abstract: Converting 2D drawings and blueprints into 3D models is an important process used in fields such as computer-aided design, architecture, engineering, and digital visualization.
A 2D drawing mainly shows the shape and dimensions of an object, but it does not always provide a clear idea about its depth and overall structure. Because of this, understanding the final design can sometimes be difficult. This research focuses on developing an efficient method for converting 2D drawings, blueprints, or images into 3D models. The proposed method analyzes the 2D input, identifies important shapes, measurements, and structural details, and then uses computer-aided design and 3D modeling techniques to create the corresponding 3D model. This process can reduce the amount of manual work required for modeling and also help minimize errors. The generated 3D models make it easier to understand the structure, dimensions, and relationships between different components. The proposed approach can be useful in areas such as architecture, construction, engineering, manufacturing, and other fields that require accurate 3D visualization. Overall, converting 2D designs into 3D models can make the modeling process faster and easier while providing a more realistic and interactive view of the original design.
A 2D drawing mainly shows the shape and dimensions of an object, but it does not always provide a clear idea about its depth and overall structure. Because of this, understanding the final design can sometimes be difficult. This research focuses on developing an efficient method for converting 2D drawings, blueprints, or images into 3D models. The proposed method analyzes the 2D input, identifies important shapes, measurements, and structural details, and then uses computer-aided design and 3D modeling techniques to create the corresponding 3D model. This process can reduce the amount of manual work required for modeling and also help minimize errors. The generated 3D models make it easier to understand the structure, dimensions, and relationships between different components. The proposed approach can be useful in areas such as architecture, construction, engineering, manufacturing, and other fields that require accurate 3D visualization. Overall, converting 2D designs into 3D models can make the modeling process faster and easier while providing a more realistic and interactive view of the original design.
Abstract
AI-DRIVEN CROP DISEASE PREDICTION AND MANAGEMENT SYSTEM
Miss.Radhika Sugdev Gavhale, Asst. Prof. Arsalan A. Shaikh
DOI: 10.17148/IJIREEICE.2026.14913
Abstract: Agriculture is very important for providing food, creating employment, and supporting the economy. However, crop diseases are a major problem for farmers because they can reduce crop production, lower the quality of crops, and increase farming costs. Finding a disease at an early stage is important because it allows farmers to take the right action before the disease spreads.
Traditionally, farmers identify crop diseases by looking at the leaves and other parts of the plant. This method can take time, and the accuracy of the diagnosis often depends on the farmer’s or expert’s experience and knowledge.
To overcome these problems, the proposed AI-Driven Crop Disease Prediction and Management System uses Artificial Intelligence, Image Processing, and Machine Learning to automatically identify crop diseases. In this system, the farmer can take a picture of a crop leaf or upload an existing image through the application. The image is first processed to make it easier for the system to analyze. A trained Convolutional Neural Network (CNN) model then examines the image, identifies important patterns, and predicts the possible disease affecting the crop.
After detecting the disease, the system provides information about the disease along with suitable management and preventive measures. The main purpose of the system is to help farmers identify crop diseases quickly and take appropriate action at an early stage. It can reduce the time needed for initial disease identification and make AI-based agricultural support easier to access.
However, the accuracy of the system depends on the quality and variety of the images used for training, as well as the quality of the images uploaded by users. Overall, the proposed system can serve as a useful decision-support tool for modern and smart agriculture.
Keywords: Artificial Intelligence, Machine Learning, Crop Disease Detection, CNN, Image Processing, Deep Learning, Smart Agriculture, Disease Prediction, Crop Management.
Traditionally, farmers identify crop diseases by looking at the leaves and other parts of the plant. This method can take time, and the accuracy of the diagnosis often depends on the farmer’s or expert’s experience and knowledge.
To overcome these problems, the proposed AI-Driven Crop Disease Prediction and Management System uses Artificial Intelligence, Image Processing, and Machine Learning to automatically identify crop diseases. In this system, the farmer can take a picture of a crop leaf or upload an existing image through the application. The image is first processed to make it easier for the system to analyze. A trained Convolutional Neural Network (CNN) model then examines the image, identifies important patterns, and predicts the possible disease affecting the crop.
After detecting the disease, the system provides information about the disease along with suitable management and preventive measures. The main purpose of the system is to help farmers identify crop diseases quickly and take appropriate action at an early stage. It can reduce the time needed for initial disease identification and make AI-based agricultural support easier to access.
However, the accuracy of the system depends on the quality and variety of the images used for training, as well as the quality of the images uploaded by users. Overall, the proposed system can serve as a useful decision-support tool for modern and smart agriculture.
Keywords: Artificial Intelligence, Machine Learning, Crop Disease Detection, CNN, Image Processing, Deep Learning, Smart Agriculture, Disease Prediction, Crop Management.
Abstract
Monitoring System for Classroom Session in Skill Training
Miss. Vaishnavi D. Ghadekar, Asst. Prof. Mr. Arsalan A. Shaikh
DOI: 10.17148/IJIREEICE.2026.14914
Abstract: Skill-training classrooms need regular monitoring to make sure that training sessions are conducted properly and learners are actively participating. It is also important to maintain accurate attendance and check whether trainers are following the planned schedule. In many training centers, these activities are handled manually, which can take a lot of time and may sometimes result in incomplete or delayed information.
This research proposes a Monitoring System for Classroom Sessions in Skill Training to make the monitoring process easier and more organized. The system provides a digital way to record attendance, track classroom activities, check session progress, and prepare reports. Trainers, administrators, and authorized supervisors can use a centralized dashboard to view the required information. The system helps reduce manual record keeping and provides better visibility of classroom activities. It can also help supervisors identify problems related to attendance, session progress, or planned classroom activities at the right time. The proposed system can be developed as a web-based application using secure login, a database, dashboards, and reporting features.
Keywords: Skill Training, Classroom Monitoring, Attendance, Session Tracking, Dashboard, Trainer Monitoring, Reporting System.
This research proposes a Monitoring System for Classroom Sessions in Skill Training to make the monitoring process easier and more organized. The system provides a digital way to record attendance, track classroom activities, check session progress, and prepare reports. Trainers, administrators, and authorized supervisors can use a centralized dashboard to view the required information. The system helps reduce manual record keeping and provides better visibility of classroom activities. It can also help supervisors identify problems related to attendance, session progress, or planned classroom activities at the right time. The proposed system can be developed as a web-based application using secure login, a database, dashboards, and reporting features.
Keywords: Skill Training, Classroom Monitoring, Attendance, Session Tracking, Dashboard, Trainer Monitoring, Reporting System.
Abstract
HEALTH DATA INFORMATION & MANAGEMENT SYSTEM
Miss. Riya D. Kachare, Asst. Prof. Arsalan A. Shaikh
DOI: 10.17148/IJIREEICE.2026.14915
Abstract: A Health Data Information and Management System (HDIMS) can centralize these records and improve their organization, retrieval and controlled sharing. However, conventional information-management systems generally focus on storing and retrieving records and do not fully exploit the analytical value of structured healthcare data. This research proposes an enhanced HDIMS that integrates a machine-learning module for heart disease risk prediction. The proposed workflow includes data collection, cleaning, exploratory data analysis, feature preparation, model training, comparative evaluation and integration of the selected model into the HDIMS interface. Logistic Regression, Decision Tree, Random Forest, K-Nearest Neighbors, Support Vector Machine and XGBoost are proposed for comparison. Evaluation will use accuracy, precision, recall, F1-score, confusion matrix and ROC-AUC, with particular attention to recall because false- negative risk can be important in screening-oriented applications. Explainable AI, such as SHAP, is proposed to help communicate the factors contributing to a prediction. The literature review indicates that machine learning has substantial research support for cardiovascular risk prediction, while external validation, data quality, interpretability, privacy and clinical translation remain important challenges. The proposed contribution is therefore not the invention of heartdisease prediction itself, but the practical integration of an interpretable, reproducible predictive workflow into a health- information management platform. The output is intended as an academic decision-support and risk-estimation feature and not as a medical diagnosis.
Keywords: Health Data Information Management, HDIMS, Heart Disease Prediction, Machine Learning, Healthcare Analytics, Electronic Health Records, Explainable AI, Risk Prediction.
Keywords: Health Data Information Management, HDIMS, Heart Disease Prediction, Machine Learning, Healthcare Analytics, Electronic Health Records, Explainable AI, Risk Prediction.
Abstract
Fake Social Media Accounts and Their Detection
Miss. Gayatri R. Gatmane, Asst. Prof. Arsalan A. Shaikh
DOI: 10.17148/IJIREEICE.2026.14916
Abstract: Social media platforms allow users to communicate, share information and build online communities. Along with genuine users, however, platforms may contain fake accounts created with false identities, copied profiles, misleading information, automated activity or the intention to perform spam and other harmful activities. Detecting such accounts is therefore an important problem in social-media security and trust management. This project presents a conceptual machine-learning based approach for identifying suspicious social-media accounts using profile, activity, network and content-related features. The proposed workflow collects account information, preprocesses the data, extracts features, trains a classification model and produces a prediction such as genuine or suspicious/fake. Common machine learning methods such as Logistic Regression, Random Forest and Support Vector Machine can be evaluated using accuracy, precision, recall and F1-score. The project also considers challenges such as changing user behaviour, class imbalance, privacy and adversarial behaviour. The objective is to develop an understandable and scalable framework that can support early identification of potentially fake accounts while keeping human review in the decision process.
Keywords: Fake Account, Social Media, Bot Detection, Machine Learning, Feature Extraction, Classification.
Keywords: Fake Account, Social Media, Bot Detection, Machine Learning, Feature Extraction, Classification.
Abstract
AI-OPTIMIZED TIME-SLOT DELIVERY FOR EFFICIENT PARCEL DISTRIBUTION IN E- COMMERCE LOGISTICS
Miss. Sanika Bhika Jaware, Asst. Prof. Arsalan A. Shaikh
DOI: 10.17148/IJIREEICE.2026.14917
Abstract: Last-mile parcel delivery is critically affected by uncertain travel times, metropolitan traffic volatility, fluctuating customer availability, spatial delivery density, courier vehicle capacities, and avoidable failed handover attempts. Conventional dispatching frameworks assign rigid, excessively broad time windows and rely on static travel- time matrices, inevitably resulting in prolonged customer waiting friction, unpunctual courier arrivals, inefficient vehicle capacity utilization, and costly repeated delivery attempts. To resolve these operational bottlenecks, this paper proposes an AI-optimized time-slot delivery framework that seamlessly couples supervised machine-learning arrival prediction with constrained time-window-aware parcel assignment and combinatorial route optimization. Historical delivery telemetry is systematically extracted into multidimensional predictive features, encompassing Haversine transit distances, historical traffic congestion indices, service dwell durations, parcel priority rankings, customer preference distributions, and neighborhood delivery densities. A supervised regression model predicts expected arrival times alongside calibrated handover success probabilities for candidate delivery intervals. Subsequently, a multi-objective optimization engine assigns parcels to feasible time slots and optimal delivery tours while strictly enforcing vehicle payload boundaries, courier shift availability, committed delivery intervals, and operational transit costs. The proposed framework is empirically benchmarked utilizing a multi-zone simulated e-commerce logistics dataset. The experimental evaluations compare a conventional fixed-slot baseline against the proposed AI-assisted paradigm across on-time delivery rates, mean delivery delay, failed attempt frequency, fleet transit mileage, and time-slot utilization efficiency. Simulation results demonstrate that integrating predictive arrival modeling with combinatorial optimization elevates on- time punctuality from 86.0% to 93.4%, reduces average delivery delay by 41.2%, cuts failed deliveries by 34.6%, and shortens fleet transit distance by 11.1%. These findings confirm that decoupled predictive-optimization architectures establish a scalable, robust foundation for next-generation e-commerce parcel logistics.
Keywords: Artificial intelligence, Last-mile delivery, Time-slot prediction, Parcel delivery, Machine learning, Vehicle routing with time windows (VRPTW), Logistics optimization, Supervised regression, Punctuality.
Keywords: Artificial intelligence, Last-mile delivery, Time-slot prediction, Parcel delivery, Machine learning, Vehicle routing with time windows (VRPTW), Logistics optimization, Supervised regression, Punctuality.
Abstract
DEVELOP A DASHBOARD AIMED AT MAINTAIN CLEAN AND GREEN TECHNOLOGY
Miss. Vidya R. Ingole, Asst. Prof. Arsalan A. Shaikh
DOI: 10.17148/IJIREEICE.2026.14918
Abstract: Clean surroundings and environmentally responsible lifestyles are essential for sustainable communities. Government programmes such as Swachh Bharat Mission and Mission Life promote sanitation, waste reduction, resource conservation and responsible citizen behaviour. However, information related to these activities is often distributed across different records, portals and local reporting mechanisms. This paper proposes a Swachhta and Life Dashboard as a unified digital platform for monitoring and communicating clean and green technology practices. The proposed dashboard combines sanitation indicators, solid and liquid waste management, water and energy conservation, plastic reduction, e-waste awareness and citizen participation into a structured visual interface. Key performance indicators can be displayed through charts, status cards, trend graphs and location-based summaries. The system is designed to support administrators in identifying gaps, monitoring progress and prioritising interventions, while also giving citizens a simple view of local environmental performance. The proposed approach emphasises data quality, role- based access, periodic updates and transparent reporting. Rather than replacing existing government systems, the dashboard is intended as an integration and decision-support layer that can bring relevant Swachhta and Life indicators together. The study presents the system architecture, methodology, dashboard modules, KPI framework, expected outcomes and future scope for smart-city and rural-development applications.
Keywords: Swachhta, Mission Life, Dashboard, Clean Technology, Green Technology, Waste Management, Sustainability, Data Visualization, Citizen Participation, Digital Monitoring.
Keywords: Swachhta, Mission Life, Dashboard, Clean Technology, Green Technology, Waste Management, Sustainability, Data Visualization, Citizen Participation, Digital Monitoring.
Abstract
Vehicle-to-Grid (V2G) Enabled Renewable-Powered Microgrids: A Sustainable Approach to Carbon Emission Reduction
Suraj Chaudhary, Himanshu Tomar, Jasmine Kaur
DOI: 10.17148/IJIREEICE.2026.14919
Abstract: The increasing use of electric vehicles (EVs) raises electricity demand and questions the environmental impacts of EV charging by electricity produced from fossil-fuel sources. Integrated renewable-based microgrids with vehicle to grid (V2G) enable coordinated renewable generation, EV charging and discharging, battery energy storage and grid exchange to minimize carbon emissions. This paper proposes an energy management framework for V2G-enabled renewable-powered microgrids by which the carbon emissions are minimized, while the utilization and operational efficiency of the system are maximized. The proposed system comprises photovoltaic (PV) generation, bidirectional EV charging, battery energy storage system (BESS) and grid-connected microgrid. An energy management policy is formulated to prioritize renewable utilization, schedule EV charging during high solar generation, utilize V2G discharging during high demand, and manage battery operations in accordance with operational constraints. The performance of the proposed framework is evaluated in terms of carbon emissions, renewable utilization, grid exchange, battery operations and operational costs. Comparative scenarios of conventional grid-connected charging, renewable- powered charging and V2G-enabled renewable microgrids are defined to quantify the contribution of each technology to the system performance. The proposed framework provides a systematic basis for assessing the technical and environmental benefits of coordinated V2G operations and their contribution to sustainable and low-carbon electricity systems.
Keywords: Vehicle-to-Grid, Renewable Energy, Electric Vehicles, Carbon Emission Reduction, Energy Management, Sustainable Development, and Smart Grid.
Keywords: Vehicle-to-Grid, Renewable Energy, Electric Vehicles, Carbon Emission Reduction, Energy Management, Sustainable Development, and Smart Grid.
Abstract
Design and Performance Analysis of Battery–Supercapacitor Hybrid Energy Storage System with PSO-Based PI Controller Optimization
Himanshu Tomar, Suraj Chaudhary, Jasmine Kaur
DOI: 10.17148/IJIREEICE.2026.14920
Abstract: The combined application of lithium-ion batteries (LIBs) and supercapacitors (SCs) in hybrid energy storage systems (HESS) allows for overcoming the power density and specific energy shortcomings of each standalone storage technology for electric vehicles (EV). The paper presents a comprehensive review, modelling analysis, and assessment of LIB-SC HESS for EVs with particular attention paid to the design of system components, topological structures, control strategies, and multi-criteria storage options. The survey provides an in-depth analysis of the passive, semi-active, and active HESS configurations along with the corresponding bidirectional DC-DC converter circuits. Particular emphasis is placed on power-energy management system (PEMS) design options, which allow for improving the battery inverter performance under highly dynamic and transient EV driving conditions. In the paper, advanced PEMS solutions based on rule-based, optimization-, and machine learning-assisted control circuits are analysed in detail. The simulation and experimental results highlight the enhanced regenerative braking capability, improved battery state-of-health, and overall vehicle performance offered by the LIB-SC HESS. The key challenges and trends in component hybridization, drive cycle parameterization, converter losses, computational complexity, and control strategy optimization are also discussed. The present work provides a basis for further research into the design and development of advanced hybrid energy storage systems for EVs.
Keywords: Hybrid electric vehicle, lithium-ion battery, supercapacitor, bidirectional DC/DC converter, Hybrid energy storage system.
Keywords: Hybrid electric vehicle, lithium-ion battery, supercapacitor, bidirectional DC/DC converter, Hybrid energy storage system.
Abstract
Development AI/ML Based Models for Predictions Prices for Agri-Horicultural Commodities Such as Pulses and Vegetables (Onion, Potato, Onion)
Miss. Dhanshree G. Mahajan, Asst. Prof. Miss. Chetana M. Kawale
DOI: 10.17148/IJIREEICE.2026.14921
Abstract: Agricultural and horticultural commodity prices are subject to considerable fluctuations due to changes in supply and demand, market arrivals, seasonal variations, weather conditions, production levels, transportation costs, government policies, and other economic factors. Such price uncertainty creates difficulties for farmers in deciding when and where to sell their produce and also affects traders, consumers, and policymakers. his research focuses on the development of Artificial Intelligence (AI) and Machine Learning (ML) based models for predicting the prices of selected agri-horticultural commodities, namely onion, potato, and pulses. Several AI/ML algorithms are considered for price prediction, including Linear Regression. these models represent different approaches to learning relationships between historical market conditions and future commodity prices. Their predictive performance is evaluated using standard statistical measures such as short long- term memory (LSTM). The developed models can contribute to improved market planning, reduced price uncertainty, and more informed agricultural decision-making.
Keywords: Artificial Intelligence, Machine Learning, Price Prediction, Commodity Price Forecasting, Linear Regression, Long short- term memory(LSTM).
Keywords: Artificial Intelligence, Machine Learning, Price Prediction, Commodity Price Forecasting, Linear Regression, Long short- term memory(LSTM).
Abstract
Identification Of Algorithm from the Given Dataset Using AI/ML Techniques
Miss. Sakshi D. Mahajan, Asst. Prof. Chetana M. Kawale
DOI: 10.17148/IJIREEICE.2026.14922
Abstract: Selecting a suitable machine learning algorithm for a given dataset remains a persistent challenge due to the wide range of available algorithms and the varying statistical characteristics of real-world data. This paper presents a meta-learning based approach that recommends a suitable algorithm family for a given tabular dataset by analyzing its statistical meta-features, such as feature correlation, class imbalance, skewness, and dimensionality. Five candidate algorithms - Linear/Logistic Regression, Decision Tree, Random Forest, Support Vector Machine, and K-Nearest Neighbors - were evaluated on sixteen benchmark datasets (five real-world datasets and eleven synthetically generated datasets covering diverse characteristics such as class imbalance, noise, and high dimensionality), and a Random Forest meta-classifier was trained to map dataset meta-features to the best-performing algorithm. Using Leave-One-Out cross- validation, the system achieved a Top-1 recommendation accuracy of 31.25% and a Top-2 accuracy of 50.00%, compared to a naive baseline of 37.50%. While the small meta-dataset size limits definitive conclusions, the results and their limitations offer useful insight into the data requirements of meta-learning based algorithm recommendation systems, and feature importance analysis identified feature correlation and class imbalance as the strongest predictors of algorithm performance.
Keywords: Meta-learning, Algorithm Selection, Machine Learning, AutoML, Meta-features, Random Forest, Baseline Comparison
Keywords: Meta-learning, Algorithm Selection, Machine Learning, AutoML, Meta-features, Random Forest, Baseline Comparison
Abstract
Enhancing Rail Madad AI-Powered Complaint Management System
Mr. Kalpesh K. Manure, Asst. Prof. Chetana M. Kawale
DOI: 10.17148/IJIREEICE.2026.14923
Abstract: Rail Madad serves as the unified digital grievance redressal mechanism of Indian Railways, facilitating complaint registration across train and station operations, ticket tracking, emergency assistance, and citizen feedback. While the platform effectively consolidates multi-channel intake into a centralized monitoring portal, first-level triage remains constrained by the ambiguities of free-form descriptions, code-mixed vernacular expressions (Hinglish), contradictory metadata, and manual department assignment. This research presents an artificial intelligence-enhanced complaint management framework for Rail Madad that integrates multilingual natural language processing, automated entity extraction, urgency and sentiment assessment, image-evidence verification, confidence-aware routing, and human-in-the-loop decision controls. The architecture adopts an incremental lifecycle model, maintaining the existing institutional complaint lifecycle while introducing AI assistance across intake, triage, prioritization, departmental routing, and continuous feedback learning. Empirical baseline analysis of official Railway Board operational performance demonstrates substantial workload escalation, rising from 1,61,575 complaints in July 2022 to 2,35,394 in
(50%), security (20%), and punctuality (15%) constitute the dominant complaint categories. The proposed framework implements a confidence-calibrated routing engine governed by institutional rules and statutory compliance with the Digital Personal Data Protection (DPDP) Act 2023. By routing low-confidence or high-impact grievances to human officers while automating routine high-confidence triage, the proposed architecture establishes a transparent, auditable, and resilient path for modernizing railway grievance redressal without compromising human accountability.
Keywords: Rail Madad, Artificial Intelligence, Complaint Management, Natural Language Processing, Complaint Classification, Sentiment Analysis, Intelligent Routing, Human-in-the-Loop, Indian Railways, Grievance Redressal.
(50%), security (20%), and punctuality (15%) constitute the dominant complaint categories. The proposed framework implements a confidence-calibrated routing engine governed by institutional rules and statutory compliance with the Digital Personal Data Protection (DPDP) Act 2023. By routing low-confidence or high-impact grievances to human officers while automating routine high-confidence triage, the proposed architecture establishes a transparent, auditable, and resilient path for modernizing railway grievance redressal without compromising human accountability.
Keywords: Rail Madad, Artificial Intelligence, Complaint Management, Natural Language Processing, Complaint Classification, Sentiment Analysis, Intelligent Routing, Human-in-the-Loop, Indian Railways, Grievance Redressal.
Abstract
Reporting Uncertainty in Permutation Feature Importance: Why the Usual Dispersion Measure Is the Wrong One
Ofierohor Ufuoma Earnest
DOI: 10.17148/IJIREEICE.2026.14924
Abstract: Permutation feature importance is usually reported as one mean value per feature and ranked, even though the standard implementation produces a dispersion measure alongside every mean. This paper argues that the dispersion normally reported is the wrong quantity, and demonstrates the consequences on a real model. For a support vector regression forecasting Nigerian FDI one quarter ahead, evaluated on 13 held-out quarters, we report permutation importance three ways: the mean, the standard deviation across 30 permutation repetitions, and a bootstrap 95% confidence interval from 1,000 resamples of the test observations. The three disagree in a way that matters. A screening rule based on the ratio of standard deviation to mean, which some authors have proposed, would clear four of six features as reliable. The bootstrap intervals show that not one of the six has an importance whose sign is established, including current FDI, which is the strongest variable in the companion econometric model. The reason is that permutation dispersion measures Monte Carlo noise in the shuffling procedure, which the analyst controls through the repetition count, while the bootstrap measures sampling uncertainty over the test observations, which the analyst does not. We recommend reporting the repetition count, and reporting bootstrap intervals wherever the holdout permits.
Keywords: Permutation Importance; Uncertainty Quantification; Bootstrap; Model Interpretation.
Keywords: Permutation Importance; Uncertainty Quantification; Bootstrap; Model Interpretation.