Browse Items (14421 total)
Sort by:
-
Intelligent Analytical Framework to Improve Customer Retention in the SaaS Industry
In the software as a service (SaaS) sector, churn is a crucial indicator as it directly affects a businesss earnings, prospects for expansion, and viability over time. Because SaaS companies mostly rely on recurring income from subscriptions, high churn rates can be detrimental to their operations. Customer retention is crucial for SaaS companies as it is frequently more profitable and cost-effective than bringing on new customers. Retention expenses and efforts can be decreased by focusing on an appropriate set of customers. This study focuses on an intelligent analytical framework that uses machine learning and artificial intelligence techniques to find the ideal group of customers for a SaaS-based organization to retain. The previous papers concentrated on either classification or survival analysis to determine the probabilities of churn. A few studies used explainable AI models to improve the predictability of the model predictions. Not having a holistic prediction model and retention strategies provides the research gap for this study. The proposed methodology used feature selection models to identify the most significant drivers of churn, and the most popular predictive models, like logistic regression, random forest, support vector machine, and neural networks, are applied to the training set. The likelihood of churn is calculated by using classification models. The Kaplan-Meier estimate is used for survival analysis to determine the odds of survival based on the tenure of each account. Lastly, the prediction models interpretability is enhanced by using explainable AI models like SHapley Additive exPlanations (SHAP) and Local interpretable model-agnostic explanations (LIME). The neural networks model gave the best accuracy of 71% for the classification model, which provided the probability of churn and the likelihood of survival, has been predicted by Survival Analysis. Explainable AI models have identified the most important features that the model considers when arriving at the probability. This enabled the company to segment the data based on the probabilities of churn and survival, and the feature importance and respective retention strategies have been planned for each segment. By implementing the suggested analytical methodology, the business may determine which customers are most important to target with customer retention strategies. The Author(s), under exclusive license to Springer Nature Switzerland AG 2026. -
Biomass-derived carbonaceous materials: Synthesis and photocatalytic applications
[No abstract available] -
Flavonol based surface modification of doped chalcogenide nanoflakes as an ultrasensitive fluorescence probe for Al3+ ion
A highly selective novel fluorescent probe was prepared by using surface modified ZnS:Mn nanoparticles, functionalized with morin, a flavonol. SEM investigations of the heterostructures prepared using wet chemical precipitation technique revealed a nanoflake type of morphology. HR-TEM and powder XRD analysis confirmed the crystalline planes corresponding to Wurtzite ZnS. The functionalized nanoparticles were characterized using Raman, XPS and FTIR which confirms the binding of morin to the nanoparticles via surface coordination. The prepared probe selectively interacts with Al3+ ions which has been used as an ultrasensitive analytical tool for determination of Al3+ ions. A major advantage of the proposed method is that the other metal ions closely associated with Al3+ did not interfere with the analysis. The detection limit and the quantitation limit were found to be 0.07 nM and 0.20 nM respectively with a linear dynamic range 0.20 nM80 nM. The method was successfully applied to environmental water samples and other complex matrices. 2017 Elsevier B.V. -
Biomass-derived carbonaceous materials: Synthesis and photocatalytic applications /
Novel Applications of Carbon Based Nano-materials, 1st ed., pp.412-429, eBook ISBN : 9781003183549. -
Unleashing the Potential: How the Internet of Things Catalyzes Progress Towards the Development Goals
The world stands at a critical juncture, grappling with intertwined challenges of environmental degradation, social inequity, and economic precarity. In response, an integrated framework has been established by the United Nations in terms of Sustainable Development Goals (SDGs), for achieving a more sustainable and equitable future by 2030. These 17 ambitious goals encompass a wide range of interconnected targets, from eradicating poverty and hunger to ensuring clean water and sanitation, fostering responsible consumption and production, and combating climate change. However, realizing this ambitious agenda necessitates a paradigm shift, one that leverages technological innovation to drive transformative change. This paper explores the burgeoning potential of the Internet of Things (IoT) as a powerful enabler in achieving the SDGs. The IoT, a webbing of interconnected physical devices embedded with sensors, software, and other technologies, offers unprecedented capabilities for data collection, analysis, and real-time communication. By harnessing this data-driven ecosystem, we can cultivate a more intelligent and sustainable world. The paper examines specific applications of IoT solutions across various SDG domains. In the realm of Clean Water and Sanitation (SDG 6), for instance, IoT-enabled smart meters can monitor water usage patterns, detect leaks promptly, and optimize water distribution networks, leading to significant water conservation efforts. Similarly, in Sustainable Cities and Communities (SDG 11), IoT-powered traffic management systems can optimize traffic flow, reduce congestion, and lower emissions, contributing to cleaner air and improved urban living. Furthermore, the paper explores the role of IoT in fostering Responsible Consumption and Production (SDG 12). By embedding sensors in products and packaging, manufacturers can gain valuable insights into resource utilization and product lifecycles. This later can then be used to optimize production processes, minimize waste, and extend product lifetimes, promoting a circular economy. In conclusion, the paper posits that the IoT, when harnessed strategically and ethically, has the potential to be a transformative force in achieving the SDGs. By fostering data-driven decision making, optimizing resource use, and promoting sustainable practices, IoT solutions can lay the foundation for a more sustainable and impartial future for all. The world faces a critical juncture, grappling with environmental degradation, social inequity, and economic precarity. The United Nations Sustainable Development Goals (SDGs) offer a blueprint for a more sustainable future. IoT technology, with its interconnected devices and real-time data capabilities, is a powerful enabler. In Clean Water and Sanitation (SDG 6), IoT-enabled smart meters equipped with water quality sensors and wireless communication protocols can monitor water usage patterns, detect leaks, and optimize distribution networks. In Sustainable Cities and Communities (SDG 11), IoT-powered traffic management systems utilizing advanced sensors and data analytics can optimize traffic flow, reduce congestion, and lower emissions. For Responsible Consumption and Production (SDG 12), IoT-enabled product tracking systems can monitor resource utilization and product lifecycles. This data can inform manufacturers on optimizing production processes, minimizing waste, and extending product lifetimes, promoting a circular economy. By harnessing IoT strategically and ethically, we can foster data-driven decision-making, optimize resource use, and promote sustainable practices, laying the foundation for a more equitable and sustainable future. The Author(s), under exclusive license to Springer Nature Switzerland AG 2025. -
Potent of sales-persons, impact on the channel of distribution in lighting industry in bangalore
Its found in array of literature on the roles, functioning of the sales persons and also illuminates how these are measured on effectiveness of channel of distribution. This study made with objective for better understanding of various variables, and out of which primary factors that could be focused for effectiveness of channel of distribution in lighting industry in Bangalore from the perceptive of intermediaries. This study draws the responses from intermediaries who are pivotal force (opinion leaders) in the market, which could prove more deep understanding for strategizing the channels in the said industry. From the review of literature we streamlined the functions performed for potent of sales persons. Further analysed with vivid using various statistical tools to understand loads (Eigen value), hence, prompting with Principal Component Analysis. This study is uses all normative way to analyse of the results reframed pivotal factors, in classifying, draining out insignificant factors. By regrouping based on the array of load, we come to understand 3 vital ingredients viz., 1) intermediaries appointment criteria 2) sales training& communication 3) concern for cost and needs of intermediaries, and urging to business institutions to opt for better channel strategy. Notwithstanding, the relationship with intermediaries are charismatic in nature, and dynamics of channel strategy would and will be determinant for success of any business organisation. 2019, Institute of Advanced Scientific Research, Inc.. All rights reserved. -
AI and Machine Learning Enabled Software Defined Networks
The telecommunications industry has not been exempt from the technology sectors massive artificial intelligence (AI) and machine learning (ML) boom in recent years. Artificial intelligence (AI) and machine learning (ML) provide advanced analytics and automation that are in line with modern networking concepts like software-defined networking (SDN) and software-defined wide-area networks (SD-WAN). Work is being done to determine how AI/ML can benefit SD-WAN and to demonstrate these benefits in a real SD-WAN network using a workable example. Modern ML techniques and algorithms are the extent of AI/ML. Todays Internet is under constant threat from DDoS (Distributed Denial of Service) attacks. As the volume of Internet traffic grows, its getting harder and harder to tell whats legitimate and whats malicious. The DDoS attack was detected using a machine learning approach that makes use of a Random Forest classifier. To better detect DDoS attacks, we tweak the Random Forest algorithm. The proposed machine learning approach outperforms, as demonstrated by our results. 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. -
Enhancement in air-cooling of lithium-ion battery packs using tapered airflow duct
Temperature uniformity and peak-temperature reduction of lithium-ion battery packs are critical for adequate battery performance, cycle life, and safety. In air-cooled battery packs that use conventional rectangular ducts for airflow, the insufficient cooling of cells near the duct outlet leads to temperature nonuniformity and a rise in peak temperature. This study proposes a simple method of using a converging, tapered airflow duct to attain temperature uniformity and reduce peak temperature in air-cooled lithium-ion battery packs. The conjugate forced convection heat transfer from the battery pack was investigated using computational fluid dynamics, and the computational model was validated using experimental results for a limiting case. The proposed converging taper provided to the airflow duct reduced the peak temperature rise and improved the temperature uniformity of the batteries. For the conventional duct, the boundary layer development and the increase in air temperature downstream resulted in hotspots on cells near the outlet. In contrast, for the proposed tapered duct, the flow velocity increased downstream, resulting in improved heat dissipation from the cells near the outlet. Furthermore, the study investigated the effects of taper angle, inlet velocity, and heat generation rate on the flow and thermal fields. Notably, with the increase in taper angle, owing to the increase in turbulent heat transfer near the exit, the location of peak temperature shifted from the exit region to the central region of the battery pack. The taper-induced improvement in cooling was evident over the entire range of inlet velocities and heat generation rates investigated in the study. The peak temperature rise and maximum temperature difference of the battery pack were reduced by up to 20% and 19%, respectively. The proposed method, being effective and simple, could find its application in the cooling arrangements for battery packs in electric vehicles. 2023 Y?ld?z Technical University. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/). All Rights Reserved. -
Machine Learning Based Depression Prediction Using Gradient Boosting Algorithm
Depression is one of the major diseases, more than one million people are facing this issue. To achieve the best results possible, it is essential to monitor and intervene when needed regularly. While there are many ways to observe the mental well-being of an individual in a workplace environment, AI has the potential to enhance the accuracy, efficiency, and speed when it comes to diagnosing any issues. This study focuses on developing an ML system for distinguishing symptoms of depression among individuals in the workplace. The dataset comprises detailed information on the signs and symptoms of depression among individuals, it particularly focuses on the observed negative consequences at the workplace, physical health issues and their negative consequences, treatment. In this experimental process two main machine learning algorithms were used, the Random Forest and Gradient Boosting algorithm. Both the algorithms have an overall accuracy of 82%, but based on maximization of the overall performance, the Gradient Boosting model is slightly better than the Random Forest. Furthermore, our exploration demonstrates overall performance like character fashions, signaling promising prospects for sturdy and correct depression analysis class systems. This study highlights the power of machine learning that could revolutionize depression care by identifying mental health problems early. The Author(s), under exclusive license to Springer Nature Switzerland AG 2026. -
Gold vs Gold Exchange Traded Funds: An Empirical Study in India
This study aim of this is to estimate the relationship between gold and Gold Exchange Traded Fund (ETF) and the performance of Gold ETFs in India by using various statistical models. The data for the study covers a period of three years from 2015 to 2018. The data was collected from the National Stock Exchange database and other sources. The outcome of this study was to find out whether there is a relationship between gold and Gold ETFs. It was found out that Gold ETFs has more returns than the physical gold; Axis ETF performed the best among those Gold ETFs selected for the study. This study will be beneficial for the market researchers and investors who find the best opportunities in the Gold ETFs. 2019 EA. All rights reserved. -
A complete person re-identification model using Kernel-PCA-based Gabor-filtered hybrid descriptors
Person re-identification is a challenging problem in computer vision. Lots of research interest is observed in this area over the past few years. A model for complete person re-identification can prove useful in this direction. Use of convolutional neural networks for pedestrian detection can improve the accuracy of detection to a larger extent. Deriving a descriptor which is invariant to the changes in the illumination, background and the pose can make the difference in the re-identification process. The predominant part of our work focuses on building a robust descriptor which can tackle such challenges. We have concentrated on building a descriptor by employing appearance-based features extracted both at local and global levels. Further, the dimensionality of the descriptor is reduced using kernel PCA. Distance metric learning algorithms are used to evaluate the descriptor on three major benchmark datasets. We propose a complete person re-identification system which involves both pedestrian detection and person re-identification. Major contributions of this work are to detect pedestrians from surveillance videos using CNN-based learning and to generate a kernel-PCA-based spatial descriptor and evaluate the descriptor using known distance metric learning methods on benchmark datasets. 2018, Springer-Verlag London Ltd., part of Springer Nature. -
Multi-frame twin-channel descriptor for person re-identification in real-time surveillance videos
Automatic re-identification of people entering the camera network is an important and challenging task. Multiple frames of the same person will be easily available in surveillance videos for re-identification. Dealing with pose variations of the person in the image and partial occlusion issues is major challenge in single-frame re-identification process. The use of more frames from the surveillance videos can generate robust descriptor to tackle issues of pose variations and occlusion. In this paper, we have emphasized on using multiple frames from the same video to generate a multi-frame twin-channel descriptor. The work deals with building a spatial-temporal descriptor which takes advantage of the twin paths to extract features of the person image. Mahalanobis distance metric learning algorithms is used for matching and evaluation. Our descriptor is evaluated on two benchmark datasets and found to surpass the performance of the existing methods. 2017, Springer-Verlag London Ltd. -
Person re-identification using part based hybrid descriptor
Real time person re-identification systems require robust descriptors for useful feature extraction. This paper focuses on a novel descriptor which can efficiently re-identify persons in varied views and change in illumination. The descriptors detect the features by dividing the person image into multiple parts. We use a combination of local and global feature descriptors to form a reliable descriptor. Performance evaluation is done on a benchmarking dataset. 2016 IEEE. -
SECURE VIDEO SURVEILLANCE SYSTEMS FOR PERSON REIDENTIFICATION USING ELLIPTIC CURVE CRYPTOGRAPHY
This chapter explores the challenges associated with person reidentification in nonoverlapping multi-camera surveillance setups, considering the wide spread use of video surveillance in public spaces. Automated techniques are crucial to handle the large amounts of data from video surveillance. The topic focuses on person detection and reidentification, addressing challenges like changes in size, pose, and background. Various methods for person reidentification are explored, emphasizing the importance of soft biometric details for accurate identification. Elliptic curve cryptography (ECC) is considered as a secure method for ensuring privacy and data integrity in surveillance systems. ECC's efficiency is highlighted in comparison to RSA, showcasing its ability to provide equivalent security with shorter key lengths, reducing computational requirements. The study employs adaptive background detection, Kalman filtering for multiple object tracking, and a CNN-based deep learning model for pedes trian image classification. Encryption using ECC secures the transmitted data, and at the receiver end, the images undergo decryption, classification, and feature extraction for person reidentification. 2026 by Apple Academic Press, Inc. -
ZnO:Al thin films from (Al2O3)x(ZnO)(1-x) powder targets by magnetron sputtering
In the present work we prepared Aluminum doped Zinc Oxide (AZO) thin films from powder targets. Various concentrations (W/W percentages) of Al2O3 such as1%, 2%, 3%, 4%, 5%, 6%, 7% and 8% were mixed in ZnO powder and made in the form of a 3 inch disc target. These ceramic targets are sputtered in RF magnetron sputtering unit for the deposition of AZO thin films. Optical and electrical properties are analyzed to get an optimized percentage of mixing for achieving high transparency and low resistivity. At Al2O3 percentage of 3% there is a considerable decrement in the resistivity, and at 7% there is a considerable decrease in the optical transmittance. Mobility and carrier concentration are increasing with Al2O3 percentage. Bandgap of the films is observed to be decreasing with increasing the Al2O3 percentage. 2021 Elsevier Ltd and Techna Group S.r.l. -
Investigation of GaSb (p+) Pocket doped GaSb/Si Vertical TFETs for High-Frequency Analog Circuits
Abstract: This paper reports on the design, modelling, and performance analysis of a GaSb/Si Heterojunction Vertical Tunnel Field-Effect Transistors (HVTFETs), employing band-to-band tunneling (BTBT). The device structure includes a p+-GaSb source and an intrinsic Si-channel/drain, forming a heterojunction that enhances tunneling efficiency due to the staggered band alignment. The obtained ON and OFF currents are 1 105 and 1 1018 A/?m. The saturation drain current (IDSat) rises with gate voltage, measured as: 2.7 108 A for VG = 0.5 V, 3.4 108 A for VG = 0.6 V and 3.9 108 A for VG = 0.7 V. The off-state current (IOFF) is very low (~1019 A) for all the VG values, indicating effective suppression of leakage current. The derived gm values for gate voltages of 0.5, 0.6, and 0.7 V are 5.4 105, 5.5 105, and 5.6 105 S, respectively, indicating effective gate control and transconductive efficiency for signal amplification. The combination of these characteristics would enable high fT and fmax, making the device suitable for broadband and millimeter-wave applications in the radio frequency (RF). The suggested GaSb/Si heterojunction vertical TFET has commendable analog/RF attributes, featuring a peak cutoff frequency (fT) of 8.91 GHz and a maximum oscillation frequency (fmax) of 5.8 GHz. The device demonstrates an intrinsic gain (Av) of 11.2 and a gain-bandwidth product (GBW) of 99.79 GHz, indicating substantial promise for RF front-end applications, including low-noise amplifiers, mixers, and voltage-controlled oscillators, as well as energy harvesting applications. Pleiades Publishing, Ltd. 2026. -
Analyzing and optimizing the usability of website access
The world wide web (WWW) plays a significant role in information sharing and distribution. In web-based information access, the speed of information retrieval plays a critical role in shaping the web usability and determining the user satisfaction in accessing webpages. To deal with this problem, web caching is used. The problem with the present web caching system is that it is very hard to recognize webpages that are to be accessed and then to be cached. This is forced by the fact that there are broad categories of users and each one having their own preferences. Hence, it is decided to propose a novel approach for web access pattern generation by analyzing the web log file present in the proxy server. Further, it tries to propose a novel hybrid policy called popularity-aware modified least frequently used (PMLFU) that best suits for the current proxy-based web caching environment. It combines features such as frequency, recency, popularity, and user page count in decision-making policy. The performance of the proposed system is observed using real-time datasets, empirically using IRCACHE datasets. 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. -
Usage data for predicting user trends and behavioral analysis in E-commerce applications
Reviewing and buying the right goods from online websites is growing day by day in todays fast internet environment. Numerous goods in the same label are available to consumers. It is thus a difficult job for consumers to pick up the correct commodity at a decent price under different market conditions. Therefore, it is important for owners of online shopping websites to better understand their customers needs and offer better services. For these reasons, the access log documented a vast amount of data related to user interactions with the websites. This access log therefore plays a key role in predicting user access trends and in recommending the best product to consumers. This research work therefore focuses on one such methodology for evaluating the pattern and behavioral analysis of users in e-commerce websites. Copyright 2021, IGI Global. -
An intelligent web caching system for improving the performance of a web-based information retrieval system
With an increasing number of web users, the data traffic generated by these users generates tremendous network traffic which takes a long time to connect with the web server. The main reason is, the distance between the client making requests and the servers responding to those requests. The use of the CDN (content delivery network) is one of the strategies for minimizing latency. But, it incurs additional cost. Alternatively, web caching and preloading are the most viable approaches to this issue. It is therefore decided to introduce a novel web caching strategy called optimized popularity-aware modified least frequently used (PMLFU) policy for information retrieval based on users' past access history and their trends analysis. It helps to enhance the proxy-driven web caching system by analyzing user access requests and caching the most popular web pages driven on their preferences. Experimental results show that the proposed systems can significantly reduce the user delay in accessing the web page. The performance of the proposed system is measured using IRCACHE data sets in real time. Copyright 2020, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. -
Automated lung cancer T-Stage detection and classification using improved U-Net model
Lung cancer results from the uncontrolled growth of abnormal cells. This research proposes an automated, improved U-Net model for lung cancer detection and tumor staging using the TNM system. A novel mask-generation process using thresholding and morphological operations is developed for the U-Net segmentation process. In the pre-processing stage, an advanced augmentation technique and contrast limited adaptive histogram equalization (CLAHE) are implemented for image enhancement. The improved U-Net model, enhanced with an advanced residual network (ARESNET) and batch normalization, is trained to accurately segment the tumor region from lung computed tomography (CT) images. Geometrical parameters, including perimeter, area, convex area, solidity, roundness, and eccentricity, are used to find precise T-stage of lung cancer. Validation using performance metrics such as accuracy, specificity, sensitivity, precision, and recall shows the proposed hybrid method is more accurate than existing approaches, achieving a staging accuracy of 94%. This model addresses the need for a highly accurate automated technique for lung cancer staging, essential for effective detection and treatment. 2024 Institute of Advanced Engineering and Science. All rights reserved.