Browse Items (16488 total)
Sort by:
-
-
A Hybrid Deep Learning and Ensemble Framework for Real-Time Cyclone Path and Intensity Prediction in Disaster-Prone Regions
Predicting the path and strength of cyclones involves significant issues in meteorology, as mistakes can greatly affect disaster management and evacuation strategies. Current models frequently encounter difficulties in achieving accurate real-time forecasting, particularly in representing complicated spatial-temporal dynamics of cyclones. The proposed study presents an innovative hybrid architecture combining deep learning and ensemble methods, using convolutional layers, LSTM units, and a gradient boosting meta-learner to improve prediction efficiency. The system was trained and verified utilising multi-year cyclone datasets obtained from Kaggle, which included atmospheric and oceanic factors. The model architecture attained exceptional accuracy, with a track error of 28 km, a mean absolute error (MAE) of 3.2 hPa for pressure, 4.5 km/h for wind speed, and a root mean square error (RMSE) of 35.4 km. The suggested approach consistently outperformed baseline models, including ConvLSTM, GRU, and XGBoost, across all critical criteria. The deployment in real-time was enabled by a containerised, low-latency API that can integrate with disaster early warning systems. This research enhances cyclone forecasting by offering a scalable, precise, and operationally feasible solution for disaster-prone areas, demonstrating practical superiority over current methodologies. The results highlight the capability of hybrid AI models to improve the accuracy and dependability of meteorological forecasts. 2025 IEEE. -
Reframing Accountability for Human Trafficking Along the India-Bangladesh Border: A Securitization and Fragmented Governance Approach
This research investigates the persistence of human trafficking along the IndiaBangladesh border through the combined theoretical lenses of securitization and fragmented governance. While existing literature and policy frameworks often approach trafficking as an issue of migration management, social welfare, or criminal justice, this study argues that such interpretations inadequately address the structural and political dynamics that shape state responses. The research posits that the framing of trafficking as a national security threatrather than a social or human rights concernhas produced a system where accountability is diffuse, bureaucratic, and performative rather than substantive. Copyright 2026, IGI Global Scientific Publishing. Copying or distributing in print or electronic forms without written permission of IGI Global Scientific Publishing is prohibited. Use of this chapter to train generative artificial intelligence (AI) technologies is expressly prohibited. The publisher reserves all rights to license its use for generative AI training and machine learning model development. -
Analyzing the synchronization and causality between Indias financial and business cycle: empirical evidence and policy insights
Purpose This research aims to develop an aggregate financial cycle for India and understand its interrelationship with the business cycle. To study this relationship, the research focuses on examining the level of synchronization, comovement and leadlag relationship between the aggregate financial cycle and the business cycle of India. Design/methodology/approach The study uses principal component analysis and wavelet transform analysis to develop the aggregate financial cycle and understand its time-frequency characteristics, respectively. Then the study undertakes a three-step econometric analysis to measure the various aspects of the relationship between the financial and business cycles. Findings The study found that the aggregate financial cycle and the Indian business cycle have long-term equilibrating relationships. The comovement and the degree of synchronization between the two cycles are moderate, which shows that the relationship between them is relatively dynamic. Further, the leadlag relationship indicated that the financial cycle often leads the business cycle and not vice versa. Originality/value The research stands out as one of the few works to capture multiple dimensions of the financial market into a single aggregate financial cycle to present a broader picture of an emerging market setting, such as India. This study adds to the literature by systematically investigating the relationship between financial and business cycles over the short-, medium- and long-term horizons. Emerald Publishing Limited -
Constructing an aggregate financial cycle for India: an analysis of key financial indicators
Purpose This paper aims to develop a financial cycle for India by aggregating key financial indicators from different sectors of the financial system and then comparing it with the business cycle to understand the macroeconomic implications. Design/methodology/approach The research uses various econometric tests to analyze the relationship among the financial indicators before applying the principal component analysis and filtering technique to extract the aggregate financial cycle. After this, the dynamic conditional correlation generalized autoregressive conditional heteroscedasticity model is applied to understand how it interacts with the economy. Findings The result of this study creates an aggregate financial cycle that effectively reflects the most influential components of the financial system, empowering policymakers to measure and safeguard the stability of the financial system and economy. Originality/value The authors contribution lies in the systematic integration of key indicators into a comprehensive aggregate financial cycle for India, which has not been thoroughly explored in existing research. This study also emphasizes the significance of banking sectors and other financial intermediaries that were undermined in the existing financial cycle studies. 2025 Emerald Publishing Limited -
Removal of Artifacts from Electroenchaphalography Signal using Multiwavelet Transform
The signal from the brain can be recorded using Electroenchaphalography (EEG). The proposed work summarizes a unique method which is used for the removal of mixed artifacts presented in the electroencephalography signal during the acquisition process. Artifacts comprises of various bio-potential unit such as electrooculogram, electrocardiogram, and electromyogram. These artifacts are referred as a noise sources which is responsible for the complexity of the EEG signal. The artifacts obtained from the EEG signal leads towards improper diagnosis of pathological conditions. The EEG signal which is obtained from the brain is the multi-dimensional signal with the various statistical properties. Time consumption of the EEG signal is not reproducible due to the biological properties of the signal. The information of the EEG signal consists of the data of the neuron levels which is collected for every millisecond with the temporal resolution scale. In account of special cases, EEG signal contains noise and artifacts where information is collected using the extraction of signals. To obtain the information of the artifacts the proposed technique is used to maintain higher accuracy in the extraction process. The proposed technique consists of multiwavelet transform to remove the artifacts from the input EEG signal. In the proposed multiwavelet transform, the signal which consists of noisy features can be decomposed using GHM and thresholding technique. This experimental analysis shows the removal of artifacts from the EEG signals. The pathological conditions are removed which leads to the increase in the accuracy of the system. Also, this research findings shows that the proposed multiwavelet transform based approach outperforms significantly with respect to conventional approaches. The reconstructed EEG signal has the lesser reliability range which is measured in-terms of signal to noise ratio and power spectral density. Published under licence by IOP Publishing Ltd. -
Proficient technique for satellite image enhancement using hybrid transformation with FPGA
Visual quality of images is improved by digital techniques for the improvement of photographs. The main purpose of image improvements is to process an image to make the output more desirable for a particular use than the original image. This paper proposes a new approach, which improves the picture of the satellite by the use of the SVD DWT concept, the Gaussian transformation DWT and multiwavelet transformation. This suggested approach would convert and approximate the single-colour value matrix of the low-flowing sub-band into one low-frequency and 15 high-frequency sub-bands, and then recreate the improved picture using the inverse transformation. In terms of technical criteria as PSNR, RMSE and CC, this approach can have higher quality and quantitative performance. This paper introduces strategies for improving hardware images using a programmable door array in real-time (FPGA). The suggested algorithm is implemented successfully with Xilinx ISE, MATLAB and ModelSim on different scale satellite images in Verilog HDL. In this article, these algorithms should be simulated and implemented using Verilog HDL. The Spartan-3E from Xilinx is the unit chosen here. 2021 IEEE -
Diagnosis of Retinal Disease Using Retinal Blood Vessel Extraction
The eye is one of the important organs of the human body. In recent times, major parts of the eye are damaged due to various retinal diseases. Major diseases related to the retina are glaucoma, papilledema, retinoblastoma, diabetic retinopathy, and macular degeneration. These diseases can be detected using image processing techniques. These diseases can cause damage to the eye; hence the early diagnosis can prevent the loss of vision. Thus the early stage of rectification may lead to smaller damage than the risky ones. By extracting the blood vessels, various retinal diseases can be identified, and the severity of the disease can also be identified. Some of these diseases in the retina will occur due to hypertension, blood pressure, and diabetics. Thus, the tear in the blood vessels leads to the loss of visuality in human beings. The proposed work consists of image processing techniques such as segmentation, feature extraction, and boundary extraction which lead to the identification of various retinal diseases with a certain level of accuracy, sensitivity, and specificity by using image processing techniques. The training and testing of retinal images are carried out by using the artificial neural network (ANN) classifier for glaucoma detection and support vector machine (SVM) classifier for detecting diabetic retinopathy. 2021, Springer Nature Switzerland AG. -
Equalization of Finite-Alphabet MMSE for All-Digital Massive MU-MIMO mm-Wave Communication
For more than twenty years, growing the performance and efficiency of wireless communications systems using antenna arrays has been an active field of study. Wireless networks with multiple-input multiple-output are also part of the current norms and are implemented around the world. Access points or BSs with comparatively few antennas are used for standard MIMO systems, and the resulting increase in spectral efficiency was relatively modest. A Multiple-Input Multiple-Output platform's capacity is researched where the transmitter outputs are processed and quantified by a set of limit quantizes through an analogue linear combining network. The linear mixing weights and cutoff levels are chosen from with a collection of possible combinations as a function of the transmitted signal. Millimetre-wave networking requires optimum data transmission to various computers on same moment network in combination with large multi-user actually massive. In order to guarantee efficient data transmission, the heavy insertion loss of wave propagation at su ch a faster speed needs proper channel estimation. A new channel estimation algorithm called Beam space Channel Estimation is suggested. From a set of possible configurations, the capacity of a massive stream from which antennas signals are handled by an analog channel as a part of the channel matrix, linear mixture weights and frequency modulation levels are selected. Probable implementations of specific analogue receiver designs for the combined network model, such as smart antenna selection, sign antennas output thresholding or linear output processing. To demonstrate the effectiveness of BEACHES in service and have FPGA implementation results, we are developing VLSI architecture. Our results show that for large MU-MIMOs, mm-wave communications with hundreds of antennas, specially made denoising can be done at maximum bandwidth and in an equipment format. Published under licence by IOP Publishing Ltd. -
An Improvised Mechanism for Optimizing Fault Detection for Big Data Analytics Environment
In the applications of fault detection, the inputs are the data reflected from health state of the observed system. A major challenge to finding errors is the nonlinear relationship between the data. Big data has other drawbacks, and the volume and speed with which it is generated are reflected in the data streams themselves. In this paper, we develop a deep learning model that aims to provide fault detection in big data analytics engine. This investigation develops an approach for fault detection in large datasets using unsupervised learning. In this research, an unsupervised method of learning is developed specifically for the task of classifying large datasets. To discover regular textual patterns in large datasets, this research use data visualization methods. In this virtual environment, we employ an unsupervised learning method of machine learning that does not require human oversight. Instead, the system should be allowed some leeway to work and find things on its own. The unsupervised learning approach utilizes data that has not been tagged. In contrast to supervised learning, this approach can handle complex tasks. 2024, Ismail Saritas. All rights reserved. -
Liquid gold: assessing groundwater quality at the historic Kolar gold fields, Karnataka, India
To assess ecological sustainability and resilience, it is necessary to periodically examine various ecological properties in areas with high pollution and contaminant risks. Kolar Gold Fields (KGF) in Kolar, Karnataka, showcases one of India's most contaminated zones because of the extensive gold mining and its lingering effects. In KGF, the quality of groundwater has been severely reduced as there exist extensive mining tailings, locally referred to as cyanide dumps, which have been neglected for several preceding years without proper disposal strategies. The current approach focuses on the water pollution caused by heavy metal deposits in the KGF region. Groundwater samples were sampled from Oorgam, an abandoned region in KGF, and subsequently filtered for water quality examinations. The investigation documented concentrations of several metals, including cadmium (0.068 0.0024 ppm), lead (0.288 0.0016 ppm), nickel (0.058 0.0047 ppm), and chromium (0.23 0.0235 ppm) and have met the standard specifications in accordance with World Health Organization (WHO). Prominent pH disparity was documented amongst the experimental samples, with a detectable pH drop in the aqua-purified water in comparison to the positive control. The test results imply that the water samples collected from KGF remain unpotable for consumption or irrigation due to the persistence of high levels of heavy metal concentration. This study underscores the urgent requirement for a remedial approach to ensure water safety for drinking and irrigation in the area. 2025 Brawijaya University. All rights reserved. -
Liquid gold: assessing groundwater quality at the historic Kolar gold fields, Karnataka, India
To assess ecological sustainability and resilience, it is necessary to periodically examine various ecological properties in areas with high pollution and contaminant risks. Kolar Gold Fields (KGF) in Kolar, Karnataka, showcases one of India's most contaminated zones because of the extensive gold mining and its lingering effects. In KGF, the quality of groundwater has been severely reduced as there exist extensive mining tailings, locally referred to as cyanide dumps, which have been neglected for several preceding years without proper disposal strategies. The current approach focuses on the water pollution caused by heavy metal deposits in the KGF region. Groundwater samples were sampled from Oorgam, an abandoned region in KGF, and subsequently filtered for water quality examinations. The investigation documented concentrations of several metals, including cadmium (0.068 0.0024 ppm), lead (0.288 0.0016 ppm), nickel (0.058 0.0047 ppm), and chromium (0.23 0.0235 ppm) and have met the standard specifications in accordance with World Health Organization (WHO). Prominent pH disparity was documented amongst the experimental samples, with a detectable pH drop in the aqua-purified water in comparison to the positive control. The test results imply that the water samples collected from KGF remain unpotable for consumption or irrigation due to the persistence of high levels of heavy metal concentration. This study underscores the urgent requirement for a remedial approach to ensure water safety for drinking and irrigation in the area. 2025 Brawijaya University. All rights reserved. -
Secure and analytical agile framework for software continuous delivery /
Patent Number: 201741040903, Applicant: Anwin Varghese.
IT world thrives on quality software products that helps business sustain and grow. These software products help reduce the effort in time & bring in better economic efficiency making these products highly dependable. The challenge of having quality product releases frequently is not quite an easy task as it sounds. To meet this challenge of producing reliable software products, the IT managers & leads need a team of dedicated developers, system programmers, testers along with a highly efficient process in place. -
A system for human face detection and recognition using feature fusion and a method thereof /
Patent Number: 202141031566, Applicant: Manjunatha Hiremath.
Biometric systems have become a vital role in the process of authenticating an individual based on physical or behavioral features/ traits of human beings. Biometric systems are categorized into two types namely Physiological and Behavioral systems. Face recognition, Fingerprint, Iris recognition, Hand geometry, and DNA fingerprint traits are considered as physiological biometrics which are essentially fixed and are relatively stable whereas voice recognition, signature and keystroke recognition are considered behavioral biometrics that can vary over a period of time due to some factors like aging, mood and behavior of the person. -
Unstructured data extraction system using multi head attention and a novel language model /
Patent Number: 202141056398, Applicant: K. P. Kavitha.
A system 100 for Offline handwritten text recognition (HTR) of a scanned handwritten text input image leveraging Modern Deep Recurrent Neural Network (RNN). System 100 comprises (RNN) is proposed with the help of the present's embodiments disclosure (RNN). A cursive eliminated handwritten text image is mapped to a multi-head attention-based sequence-to-sequence learning applying the beam search technique and employing an RNN-based variable-length encoder-decoder architecture. -
A Relative Reference Responsive LRU based Page Replacement Algorithm for NAND Flash Memory
The acceptance of NAND flash memories in the electronic world, due to its non-volatility, high density, low power consumption, small size and fast access speed is hopeful. Due to the limitations in life span and wear levelling, this memory needs special attention in its management techniques compared to the conventional techniques used in hard disks. In this paper, an efficient page replacement algorithm is proposed for NAND flash based memory systems. The proposed algorithm focuses on decision making policies based on the relative reference ratio of pages in memory. The size adjustable eviction window and the relative reference based shadow list management technique proposed by the algorithm contribute much to the efficiency in page replacement procedure. The simulation tool based experiments conducted shows that the proposed algorithm performs superior to the well-known flash based page replacement algorithms with regard to page hit ratio and memory read/write operations. 2021, Webology. All Rights Reserved. -
Review on developments in nand flash page replacement algorithms
The non volatility, low power consumption and high density of NAND flash memories, made them an inevitable part of electronic industry. Due to the high wear out nature exhibited by flash systems, the algorithms used for page replacement in traditional memory systems are not suitable for flash page replacement. Along with the objective to maintain high hit rate, flash page replacement algorithms should aim at decreasing the page write count and maintain wear levelling. This paper presents major algorithms proposed for flash memory page replacement. The major contribution of this work is a relative study on various strategies, performance matrices and evaluation tools used for flash page replacement algorithm. The study shall help the researchers to identify the pros and cons of various flash page replacement algorithms, to identify the major gaps in between and to identify some commendable tools that can be used for flash page replacement algorithm evaluation. The gaps identified need to be addressed seriously in the near future. 2020, Engg Journals Publications. All rights reserved.