Predictive Analytics Research Papers - Academia.edu (original) (raw)

2025, Deep Science Publishing

To compete in the digital economy, retail industry firms must supply the right product, in the right quantity, at the right time, to the right place, and at the right price. Supply chains must be customer-centric and align with customer requirements. They are challenged to shorten cycle lead times as customer demand shifts to the just-in-time ordering, while increasing the inventory to meet service objectives in the face of increased demand volatility and unpredictability. Supply chain activities, especially warehousing and inventory management, are crucial to reduce total business costs, and technology-enabled decision tools are required to optimize these activities. Online channels are the most disruptive factor in today's retail environment, and consumers are increasingly using these channels to select their products while wanting to avoid delays in order fulfillment. This requires additional demand for fulfillment services from operation of distribution and retail branches with complex inventory policies (Choi et al., 2018; Duan et al., 2019; Ghosh et al., 2021). The disruptions have needed collaborative arrangements to reconceptualize the nature of core business, which involves the reconsideration of the product and service mix, the appropriate distribution channel, and the collaborative resources and capabilities required to support the service and product development process over the long term. Retail supply chains are now designed and managed by networks that represent more than just the flow of logistics in the form of distribution channels and suppliers, and new paradigms in performance assessment and resource planning are needed. Profit recovery is achieved through accurate prediction replacement part demand, particularly in demand spikes. Traditional demand forecasting methods use past demand history, but sales of many products exhibit seasonal, cyclical, or trend-stabilizing patterns, which are difficult to model. Time series, autoregressive statistical methods have been unable to provide accurate forecast results in terms of MSE or MAPD

2025, ABCD Index

Predictive maintenance (PdM) has emerged as a vital component of industrial asset management, and the integration of Artificial Intelligence (AI) and Machine Learning (ML) is revolutionizing its effectiveness, especially in natural gas... more

2025

Data has become a consumer asset in the digital age, requiring innovative solutions for storage, processing and analysis. Cloud computing has emerged as a transformative technology that enables organizations to better manage large amounts of data and improve business flexibility by reducing costs This paper presents case studies of intelligent business applications (IBAs) such as customers relationship management (CRM) and enterprise resource management (ERP) systems -Examine how they improve data management Studies show that cloud integration gives organizations real-time data access, scalability, and advanced analytics capabilities, and enables them enable faster and more accurate decision-making Despite the obvious benefits, challenges remain such as data protection, privacy, and compliance with regulations such as the GDPR and CCPA. This study identifies these challenges and proposes solutions, including improved encryption, access control, and hybrid cloud models. The findings suggest that future advances in cloud technologies such as AI, blockchain, and edge computing will further increase the efficiency, security, and scalability of cloud-based enterprise data.

2025, Kipkoech Ezrah

In this paper, we examined the risk character of the NSEASI index across 10 years (January 1, 2013 -August 31, 2023), consisting of around 2,590 valid trading days following intensive cleaning and outliers adjustment of the data. A daily log return was calculated and shown as a high-risk, low-reward market, with average log returns of 0.0018 and an 11.73% daily volatility. It had extremely high kurtosis (328.199) and almost zero skewness (0.009), implying that the distribution of returns was very skewed to extremes and was not skewed. The characteristic function-based Value at Risk (VaR) model was applied in a stochastic volatility system to rectify the flaw of traditional risk models in the face of this heavy-tailed behaviour. Realistic stochastic dynamics of volatility of returns were obtained using parameter estimation using the method of moments. Comparative analysis using Delta, Delta-Gamma, and Monte-Carlo simulation techniques revealed that the fat-tailed behaviour of the return distribution was better captured when using the CF-based and Monte-Carlo-based approaches. The estimates of VaR at the 5% and 1% confidence levels based on CF (2.80 and 5.10) were significantly higher than those of the Delta and the Delta-Gamma method, which underestimated tail risk. It provides formal backtesting via the Kupiec and Christoffersen tests. It performs a sensitivity analysis and discusses policy implications in the context of financial regulation and corresponding portfolio risk management. We would conclude that CF-based VaR is a more practical and theoretically-grounded alternative to more common methods, in non-Gaussian settings that characterize emerging markets; nevertheless, our findings demonstrate the shortcomings of standard Gaussian-based models in turbulent emerging markets like Kenya. This article recommends the use of advanced stochastic methods in the field of financial risk management and regulation. Future research opportunities include introducing the dynamics of jump-diffusion processes, modeling interdependencies at the constituent level, and improving the dynamic portfolio risk estimation.

2025, Inspira-Journal of Commerce, Economics & Computer Science

In the contemporary business landscape, artificial intelligence (AI) has emerged as a transformative force across multiple industries, with India positioning itself as a significant growth center for global technology firms, as evidenced in the World Bank's 2022 "Navigating the Storm" report. Many organizations strive to maintain competitive advantage and operational efficiency; AI-driven solutions have become increasingly essential components of modern HRM practices. The implementation of AI technologies in HRM processes offers considerable benefits, including cost reduction, time optimization, and enhanced strategic resource allocation. The application of AI in recruitment, on boarding, and performance management, which substantially reduces administrative burden for HR departments. This study examines the pivotal role of AI integration within human resource management (HRM) functions, analyzing both its strategic advantages and inherent challenges. Organizations face significant challenges including high implementation and maintenance costs, data privacy concerns, and cyber security vulnerabilities. While existing research has extensively explored AI applications in recruitment and selection processes, notable research gaps remain in other critical HRM domains including performance management, career development, training initiatives, and employee retention strategies. This study investigates the multifaceted implications of AI integration across HRM functions. AI, defined as a computer science field that leverages databases for accelerated problem-solving and decisionmaking, represents a genesis technology experiencing multiple developmental cycles. As an umbrella term encompassing machine learning, deep learning, and natural language processing, AI offers significant potential for automating routine tasks, thereby reducing costs, conserving time, and minimizing human error. This research contributes to the understanding of how AI technologies can be strategically deployed to enhance HRM effectiveness while addressing implementation challenges.

2025, Deep Science Publishing

In the last decade, rapid developments in sensing technologies have generated an unparalleled quantity of data associated with industrial processes, enabling a shift from traditional quality control practices to quality aware production chains. The availability of smart sensors for condition and quality monitoring, together with advances in predictive maintenance models, can contribute to the reduction of production variability, increased end-product quality, and optimization of safety and cost effectiveness. Noting that quality control has been traditionally associated with the inspection of finished products and processes, the opportunity for the development of quality aware production processes is studied. It is emphasized that efforts should be directed towards ensuring that the necessary conditions are satisfied while the product is under construction. If actions are undertaken to control and optimize the key process variables associated with quality, then the finished products will not need to undergo expensive destructive inspection when their quality is assessed. It is recognized that the incentive for the development of real time quality control practices lies with the reduction of production costs, achieved through reduced inspection requirements and lower warranty costs. Furthermore, real-time information associated with product quality can utilize causal databases to allow companies to optimize their quality, by directing them to situations in which the establishment of product quality is compromised. Traditionally quality control would not look for insights into production improvement, while today by employing modern quality control practices, companies can fulfill today's expectations for product quality, reliability, and safety . Additionally, the operators could miss some faults leading later to product damage, which might last more than a short period of time. Because of that, the quality control policy might not be efficient if the final check is performed by humans only. In fact, it Deep Science Publishing

2025, International Journal of Engineering Science and Advanced Technology

In the age of digital transformation, social media has emerged as a powerful tool for marketers to understand, influence, and predict consumer behavior. This study develops a robust framework that leverages social data mining and probabilistic modeling to enhance audiencetargeted marketing. The research integrates user interaction data, sentiment analysis, and network structure to build predictive models capable of identifying high-value targets for online advertising. A distributed Lasso-based regression technique, coupled with Singular Value Thresholding (SVT), is employed to address issues of data sparsity and scalability. Additionally, the system uses probabilistic Bayesian networks to infer social brand reputation and user positivity, thereby enabling the dynamic ranking of products and influencers. Realworld implementation using data from platforms like Pinterest and Facebook validates the system's effectiveness in optimizing marketing campaigns. The findings highlight the importance of combining textual insights with user behavior and network dynamics to achieve precision targeting. This work offers a computational foundation for predictive marketing strategies and emphasizes the growing relevance of big data in marketing intelligence. It serves as a blueprint for businesses aiming to transform vast social data into strategic insights that drive consumer engagement and brand loyalty.

2025

Makalah ini membahas integrasi Power BI dengan algoritma Adaptive Neuro-Fuzzy Inference System (ANFIS) untuk analisis data prediktif. Power BI menyediakan platform visualisasi data yang kuat, sementara ANFIS menawarkan kemampuan pemodelan non-linier dan inferensi yang adaptif. Penelitian ini akan mengeksplorasi bagaimana kedua teknologi ini dapat digabungkan untuk menghasilkan wawasan prediktif yang lebih akurat dan mudah dipahami, khususnya dalam konteks [Sebutkan konteks/bidang aplikasi spesifik, contoh: predikisi penjualan, peramalan permintaan, analisis sentimen, dll.].

2025, Journal of Scientific and Engineering Research

This research explores the application of reinforcement learning (RL) in real-time risk management for financial stress testing. The RL agent simulates volatile market conditions, regulatory shocks, and counterparty failures to assess portfolio vulnerabilities. By continuously learning from live economic indicators and extreme events, such as pandemics and geopolitical conflicts, the system adapts and evolves, offering financial institutions a proactive approach to managing risks. This study aims to improve the resilience of financial portfolios, preventing systemic collapses in a constantly changing economic landscape.

2025, International Journal of Science and Research

As Large Language Models (LLMs) scale beyond trillion parameters, traditional Mixture of Experts (MoE) routing mechanisms face critical limitations in efficiency, load balancing, and intelligent expert selection. This paper presents a comprehensive analysis of next-generation MoE architectures specifically designed for LLM systems, addressing fundamental challenges in large-scale language model deployment. We systematically examine three transformative approaches: Mixture of Tokens (MoTs) that achieve 3× LLM training speedup through group-based token processing, LLM-powered routing that leverages language models' reasoning capabilities for intelligent expert selection, and federated MoE architectures enabling privacy-preserving distributed LLM inference. Our analysis of production LLM systems reveals cost reductions of up to 85% while maintaining 95% performance retention compared to monolithic language models. We introduce formal frameworks for capability-aware LLM routing and contextual bandit optimization tailored for language model characteristics. Through extensive benchmarking on language understanding tasks (MMLU, MT Bench, GSM8K) and real-world LLM deployments, we demonstrate that next-generation MoE systems fundamentally outperform traditional approaches in LLM scalability, adaptability, and computational efficiency. Our findings establish a technical roadmap for intelligent LLM orchestration systems with direct implications for enterprise AI deployment strategies.

2025

Outlier Detection is a Data Mining Application. Outlier contains noisy data which is researched in various domains. The various techniques are already being researched that is more generic. We surveyed on various techniques and applications of outlier detection that provides a novel approach that is more useful for the beginners. The proposed approach helps to clean data at university level in less time with great accuracy. This survey includes the existing outlier techniques and applications where the noisy data exists. Our paper defines critical review on various techniques used in different applications of outlier detection that are to be researched further and they gives a particular type of knowledge based data i.e. more useful in research activities. So where the Anomalies is present it will be detected through outlier detection techniques and monitored accordingly especially in educational Data Mining.

2025

Shariah boards (SBs) play a unique role by providing assurance on religious compliance of Islamic banks. In fulfilling this governance responsibility, SB members must exercise diligence, independence and be transparent at all times. In this study, we examine the distinct governance structure of Malaysian Islamic banks as the country is perceived to have the most developed governance framework in the Islamic world. We critically evaluate the literature on the role and function of SBs and assess the existing support mechanisms available and identify challenges in providing effective religious compliance reviews. We highlight the inherent limitations of Malaysian SBs in performing the compliance review with studies reporting the independence of this authority being compromised with SBs appearing to rubber-stamp decisions already taken by top management. Finally, we highlight serious concern regarding the quality of the religious assurance provided in annual reports of Islamic banks and...

2025, Journal of Electrical Systems

Cloud-native architectures demand dynamic scaling mechanisms to balance performance, cost, and resource efficiency. Traditional reactive scaling methods often fail to address volatile workloads, leading to over-provisioning or service degradation. This paper proposes an AI-driven predictive scaling framework leveraging time-series forecasting, reinforcement learning, and hybrid models to anticipate resource demands and optimize cloud-native systems. We present a systematic evaluation of algorithms like LSTM and Prophet, integrated with Kubernetes orchestration, to demonstrate 35-40% cost reduction while maintaining 99.9% QoS compliance. Challenges such as data noise, model explainability, and ethical implications are critically analysed, alongside future directions in federated learning and energy-aware scaling.

2025, International Journal of Innovative Research in Electrical, Electronics, Instrumentation and Control Engineering

The optimization of supply chain architecture is essential for sustainability among industries. Dynamic evolutionary events such as natural disasters, pandemics, terrorism attacks, political unrest, and cyberattacks have severe implications that disrupt the operations incurred by an unplanned event or attack and economic losses that reduce the agility and resilience of supply chains. Organizing initial inventory stock to meet uncertain customer demand is a primary action to counteract uncertainties during the onset of the crisis. Traditional supply chain models may not be effective enough considering past order and demand patterns, incorporate time-dependent fixed-up and shutdown costs, and respond slowly toward uncertain demand due to dissemination issues. Cloud computing, sensor, and network technologies, and distributed intelligent systems provide a universal and economical information and communication technology infrastructure for supply chain entities to share supply chain information, big data analytics, decision support tools, and systems through the Internet regardless of time and place. This innovative approach leads to create a consolidated Virtual Supply Chain (VSC) based on cloud models that associates and connects, in a dynamic fashion, any supply chain entity capable of storing, sharing, and processing information to support multiagent, collaborative, ubiquitous, instantaneous decision-making and recommending answers. The use of advanced cloud computing, the Internet of Things, and data analytics technologies provides the supply chain actors with capabilities of forecasting demand, managing risks, and optimizing supply chain design, planning, and operations in a collaborative, efficient, aligned, responsive, resilient approach. Considering that customers have become the driving agents and have the power to make a difference concerning final demand, this chapter proposes a demand perspective for modeling, optimizing, and managing the VSC by developing an advanced commercial agent-based simulation model.

2025, Deep Science Publishing

In 2020, humankind was awakened to the harsh reality that, despite several initial successes in different fields of physical and life sciences, we are still not fully equipped to deal with biological and healthcare-related issues and are incredibly vulnerable to unforeseen events like the COVID crisis. Rapid developments have been made in many different areas related to disease predictions and applications; however, research and development in predictive models, especially in healthcare and medical decisionmaking, have not been satisfactory and are still in a premature stage. Because of all these reasons, we have been encouraged to explore further the application of new and innovative methods in predictive healthcare and illustrate many different areas related to that. This book is intended as a first step in this direction. Health is one of the most important aspects of an individual's life as well as of a society as a whole. Over the past several years, numerous research activities around the world have been undertaken in different areas related to the field of public health. Numerous predictive models have been developed as data mining tools that can be deployed to enhance the decisionmaking capability of individuals as well as of policymakers and authority figures in the healthcare field. Specific examples include predicting the epidemic of diabetes, cardiovascular diseases, chronic diseases, COVID-19, and even predicting the number of infections and deaths caused by COVID-19. These models can assist individuals as well as authority figures in the healthcare field in planning and executing the necessary tasks in the timely management of these diseases, thus reducing the number of hospital visits and admissions and the consequent medical expenditure associated with this disease. In addition to these distinct areas, predictive models have been used in several other distinct healthcare issues (

2025, Deep Science Publishing

With the rapid advance in artificial intelligence and cloud technology, financial advising has, until recently, been considered well outside the stance of technology-founded approaches. However, in recent years, a plethora of applications capturing these new technology opportunities have sprung up. Additionally, the respect for the traditional user bases of financial advising is very high and based on years of building convincing deals. Also, the prospective partnership with the traditional value chains implies the continued existence of the consumer base (Asatryan, 2017; Chang et al., 2017; Munoko et al., 2020). With this new robust and configurable AI literature, various opportunities arise. This includes investing opportunities but also creates a problem space. Traditionally, these agents, i.e., the software applications powered by AI technology and acting as financial planners, wealth management specialists, agents, or even just assistants, have been trained, used, or prepared in particular programs. They are pre-configured for similar market conditions, risk profiles defined by or very close to their respective investments and wealth management firms, or just the tools they were designed for. Quantitative advisory, i.e., assisting or even automating the decision process in financial dealing. This includes alternatives such as investment and wealth management but also loans and consumer financial products. Particularly the quantitative part of this approach, including risk prediction, risk assessment, prediction of financial product performance, and transaction actions based on the risk and audit levels, has been very Deep Science Publishing

2025, Saira Salvi

This paper explores how Big Data Analytics is revolutionizing the retail sector. With the exponential growth of data from online and offline sources, smart retail systems are leveraging big data to improve customer experiences, inventory management, and operational efficiency. We present a study of various technologies, tools, and frameworks that are enabling this transformation along with real-world examples.

2025, Deep Science Publishing

One of the major milestones in the rise of the learning analytics field was the development and use of institutional data to construct models that leverage algorithmic modeling to assist in the prediction of students' educational outcomes. Institutions are increasingly using data to gain new insights and improve student learning outcomes with actionable insights and personalized interventions. Educational data mining focuses on methods that model and support the prediction of different actions to support students' learning and its outcomes. Predictive models have addressed questions like who is more likely to succeed or how we can help a student who is struggling. Predictive analytics can help identify attrition-related factors, thereby allowing institutions to perform proactive interventions that help students succeed. For teacher education programs, the use of learning analytics and educational data provides important support for data-driven decision-making. Data-driven approaches can lead to improvements in assessments, understanding of the course content efficacy, and comprehension challenges and obstacles that students may face within these requirements. Still, algorithms are not flawless, and ethical considerations must be considered when making data-driven decisions. Since the graduation rates have significantly improved during the last ten years, despite the selective criteria, we see the potential to focus on how these models can be used to improve graduation rates without sacrificing program admission criteria. For that, teacher education program employment data and reasonable completion periods may be critical for indicating potential for graduation.

2025

Fintech Banking: Regulatory Architecture, Revenue Models, and Systemic Economic Implications
This academic paper examines fintech banks as hybrid financial institutions that combine traditional banking services with technology-driven operational models. The analysis addresses three primary research areas: regulatory compliance challenges within existing banking frameworks, innovative revenue generation mechanisms through platform effects and data monetization, and broader macroeconomic implications for monetary policy transmission and financial inclusion.
Key Research Contributions:
Regulatory Framework Analysis - Investigates tensions between innovation velocity and prudential oversight, examining independent licensing versus partnership arrangements with established institutions, and cross-border regulatory arbitrage opportunities.
Revenue Model Innovation - Analyzes the transition from traditional intermediation-based revenue to ecosystem monetization through Banking-as-a-Service platforms, network effects, and algorithmic financial advisory services.
Systemic Risk Assessment - Evaluates technology infrastructure dependencies, cybersecurity vulnerabilities, and potential contagion pathways that differ fundamentally from traditional interbank exposures.
Economic Impact Evaluation - Examines monetary policy transmission mechanism alterations, financial inclusion effects, and market efficiency improvements through reduced transaction costs and enhanced resource allocation.
Methodology: Combines theoretical financial economics frameworks with empirical analysis of operational models, regulatory responses, and market performance data across multiple jurisdictions.
Target Audience: Financial economists, regulatory policy researchers, and banking industry analysts interested in digital transformation's impact on financial system architecture and stability.
The paper concludes that fintech banks represent a fundamental evolution in financial intermediation with significant revenue growth potential, contingent upon successful navigation of regulatory evolution and competitive market maturation.

2025, TEACHERS’ PERCEPTIONS ABOUT TEACHER EDUCATION PROGRAM IN PAKISTAN IN PERSPECTIVE OF GLOBAL EDUCATION

A descriptive research was carried out in order to pinpoint aspects related to the global viewpoint
in teacher education and offer potential recommendations for enhancement. It is imperative that
educators receive training on global concerns since they are essential to the development of a
society. Finding global components in Abdul Wali Khan University's English B Ed courses was
the study's main goal. The population was made up of all instructors at Abdul Wali Khan
University and its associated institutions who taught at the B Ed level. Thirty-six teachers were
chosen at random to make up the sample. In order to evaluate and tabulate the data, descriptive
statistical methods were used.Teachers were found to be adequately aware of the goals of the BEd
level courses. Though its usage is not encouraged, the curricula included principles of using the
Internet, multimedia, and other information technology tools in the teaching and learning process.
Women's rights, gender equality, and social justice were not included in the conceptual framework
of the selected subject's curriculum. It was observed that the courses failed to foster pupils' critical
thinking skills. It was suggested that curriculum planners make sure that their curricula don't
include any content that emphasizes blind trust in the past. Instead, teachers should help pupils
build a realistic and analytical perspective based on logic, reason, and rationality. The use of
electronic media, the internet, and other recently released educational applications must be taught
to aspiring teachers. Integrating these into the curriculum will surely help society become more
aware of global education and prepare for the challenges of the modern world.

2025, Journal of Information Systems Engineering and Management

Endometriosis is a chronic condition in which the lining of the uterus grows outside the uterus, causing pain, swelling, and fertility problems. It usually affects the ovaries, fallopian tubes, and pelvic lining, leading to severe menstrual cramps and other complications. Traditional methods for diagnosing endometriosis, such as laparoscopy and ultrasound, are often invasive, time-consuming, and can lead to delayed diagnosis. Relying on symptom-based assessment lacks accuracy, making early and affective treatment challenging. To solve the problem a novel Hyper capsule Resnet50-CNN algorithm is introduced for classifying the ovarian cysts by utilizing the ultrasound images processed datasets and applied to the image processing technique. Initially, Butterworth Filter preprocessing enhanced the details of the input data set and gave a clear view of the input dataset. Modified Watershed Segmentation algorithm (MWSA) separates follicles or cysts that specifically differentiate for selection features. An improved Recursive Bee Colony (RBC) Feature Selection algorithm is trained to identify biologically significant markers, ensuring accurate feature extraction without errors.ResNet50 with CNN architecture is a deep learning approach to extract complex features methodically hyper capsule ResNet50 contains 50 layers of network operation, which is a disappearing gradient issue that is frequently observed. RESNET 50 classification identifies ovarian cysts into three stages based on the condition: regular nodule, ovarian growth, and polycystic ovary. The accuracy is 94.15%, sensitivity 95.82%, Specificity 94.54%, FI– Score
94.89% and RMSE 84.25% measure parameters are analyzed, and the
performance Matrix obtains results.
Keywords: Endometriosis, Recursive Bee Colony (RBC), hyper capsule
Resnet50-CNN algorithm, RESNET 50 classification

2025

System (PACS) telah lama menjadi tulang punggung pengelolaan data pencitraan medis di berbagai fasilitas kesehatan. Sistem ini berfungsi sebagai repositori sentral untuk semua data citra medis, yang sebagian besar dalam format standar Digital Imaging and Communications in Medicine (DICOM). Dalam konfigurasi klasik, server PACS umumnya dioperasikan secara lokal di dalam lingkungan rumah sakit itu sendiri, seringkali dengan dukungan pencadangan data yang terbatas pada infrastruktur internal. Meskipun pendekatan ini telah terbukti memadai untuk memenuhi kebutuhan praktik klinis berskala menengah, seiring dengan pesatnya perkembangan teknologi dan volume data kesehatan yang terus meningkat, sistem tradisional ini mulai menunjukkan keterbatasannya. Tantangan utama yang muncul meliputi ketidakmampuan untuk mengatasi pertumbuhan data pasien dan citra medis yang eksponensial (Larson et al., 2021). Selain itu, terdapat kesulitan dalam memfasilitasi distribusi dan akses data lintas wilayah, yang sangat krusial dalam era kolaborasi medis modern (Liu et al., 2020). Beban pemeliharaan rutin dan kebutuhan akan pembaruan perangkat keras yang konstan juga menjadi masalah tersendiri, terutama bagi fasilitas kesehatan di daerah dengan keterbatasan infrastruktur Teknologi Informasi (TI) yang memadai (Kalra et al., 2018). Oleh karena itu, kebutuhan akan solusi yang lebih skalabel, fleksibel, dan efisien menjadi semakin mendesak. Pergeseran paradigma menuju solusi berbasis cloud computing dan integrasi kecerdasan buatan (AI) menawarkan jalan keluar yang menjanjikan. Ini tidak hanya mengatasi keterbatasan sistem PACS konvensional, tetapi juga membuka peluang baru dalam peningkatan efisiensi alur kerja radiologi, akurasi diagnostik, dan kemampuan kolaborasi lintas institusi secara signifikan.

2025, Al-Kindi Centre for Research and Development, London, United Kingdom

Personal debt in the United States has reached critical levels, creating widespread economic strain and limiting opportunities for financial mobility. This article presents a comprehensive AI-driven ecosystem designed to proactively identify financially distressed individuals and connect them with personalized debt relief resources through advanced machine learning and realtime data engineering. The framework integrates multiple AI models, including risk classification algorithms, propensity scoring systems, natural language processing for intent detection, and recommender systems for tailored program matching. Built on a scalable infrastructure utilizing Apache Kafka and Spark for stream processing, the system aggregates behavioral signals from diverse sources while maintaining privacy compliance. The multichannel engagement strategy encompasses on-site personalization, targeted digital remarketing, connected television campaigns, and direct communication channels to ensure inclusive reach across demographics. Through a structured five-phase journey from crisis identification to financial empowerment, the framework demonstrates significant improvements in program participation rates, debt reduction outcomes, credit score rehabilitation, and reduction in financial anxiety. The system's architecture enables nationwide deployment across varied populations and regions, offering a transformative solution to address economic inequality and promote sustainable financial recovery. This technological innovation represents a convergence of artificial intelligence, behavioral science, and social impact, providing a blueprint for large-scale financial wellness initiatives that serve the public good while advancing the field of applied AI in economic contexts.

2025, Dhevio Oktofitra Rahadikusuma

Disusun untuk memenuhi tugas mata kuliah Manajemen Logistik Medik dan Non Medik Rumah Sakit

2025, Zenodo (CERN European Organization for Nuclear Research)

The decline in global biodiversity is a pressing concern due to human activities, leading to millions of species at risk of extinction. East Africa is especially affected by habitat destruction, poaching, and climate change, resulting in significant losses in wildlife populations. Machine learning (ML) has demonstrated potential in identifying species, especially in camera trap images, acoustic recordings, and genetic data. However, there is a need to further explore the use of ML in identifying wildlife species in East Africa. To address this need, we developed ML classification models to identify wildlife species in East Africa. Our dataset included taxonomic features and characteristics of wildlife species from East African countries between 2018 and 2021. We used the random forest algorithm, which is suitable for complex datasets with multiple features. Our evaluation achieved an accuracy of 63.4% and a baseline score of 8.02%, showing the potential of our models in identifying wildlife species in East Africa. Our study could contribute to wildlife conservation by detecting and preventing illegal wildlife trade activities, monitoring population trends, assessing the impact of human activities on different species in East Africa, and preserving biodiversity.

2025, Advances in Consumer Research

The increasing complexity and vulnerability of modern supply chains, exacerbated by geopolitical tensions, climate variability, and fraudulent activities, highlights the need for robust AI-driven risk management solutions. This research presents a unified, data-driven framework that utilizes machine learning (ML), deep learning (DL), and reinforcement learning (RL) to enhance supplier risk resilience and optimize logistics under disruptive conditions. We use a comprehensive dataset of 1,000 supplier transactions, enriched with historical demand, weather indices, geopolitical risk scores, shipment anomalies, and financial health indicators. We apply various regression models, including Linear Regression, Random Forest Regressor, XGBoost Regressor, and Multi-Layer Perceptron, to forecast future demand and quantify supplier risk, assessing performance with metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), and R². Next, we employ Isolation Forests for real-time disruption detection, analysing features like price spike percentages, delivery delays, and sentiment scores to enable the early identification of anomalous events. To optimize dynamic routing in the face of stochastic disruptions, we design a custom Open-AI Gym environment and train a Deep Q-Network (DQN) agent that balances fuel costs, delays, and penalties for anomalies, evaluating the strategy's effectiveness through cumulative reward analyses. Finally, we built a deep neural network using a synthetic fraud dataset for transactional fraud detection, applying SMOTE for class balancing. This results in near-perfect accuracy (>99.9%), as validated by train/validation loss curves and classification reports. The integrated framework provides end-to-end supplier risk analytics, combining predictive forecasting, anomaly detection, route optimization, and fraud identification to support resilient decision-making in supply chain operations. Key evaluation metrics include MAE, MSE, and R² for forecasting; contamination rates for anomaly detection; cumulative rewards for reinforcement learning performance; and accuracy, precision, recall, and AUC for fraud classification.

2025, International Journal of Science and Research (IJSR)

As organizations become more aware of how artificial intelligence (AI) can accelerate digital transformation, it will become crucial for them to manage the unpredictable aspects of AI-related projects to ensure sustained positive outcomes. This paper presents a structured approach to assist organizations that are tasked with dealing with the inherent indeterministic nature of AI-type projects. The structured approach includes uncovering the feasibility of AI in definite organizational contexts, initial project scoping and experiment planning, designing a human-in-the-loop (HITL) and feedback framework, and developing a sustainable governance and accountability model. By utilizing a structured process, organizations can align their AI deployments with business outcomes while mitigating corporate risks associated with the probabilistic nature of AI outputs.

2025, Journal of Neonatal Surgery

The integration of Machine Learning (ML) and Convolutional Neural Networks (CNNs) has significantly advanced predictive analytics in healthcare. These technologies enable the analysis of complex medical data, facilitating early diagnosis, personalized treatment, and efficient resource allocation. CNNs, renowned for their prowess in image recognition, have been effectively applied to medical imaging tasks such as tumor detection, diabetic retinopathy classification, and organ segmentation. Simultaneously, ML algorithms, including decision trees and support vector machines, complement CNNs by processing non-image-based medical data, aiding in patient risk assessment and prognosis prediction. Despite these advancements, challenges persist, including data scarcity, class imbalance, model interpretability, and ethical concerns regarding patient privacy. This paper explores the current landscape of ML and CNN applications in healthcare prediction, highlighting their capabilities, limitations, and potential future directions.

2025, Pathfinder of Research

The accelerated progression of machine intelligence and generative AI (GenAI) is escalating the landscape of U.S. business, introducing transformative opportunities while raising critical societal questions. This review article systematically examines fourkey dimensions of AI's impact: technological innovations, practical applications, emerging challenges, and future directions. Regarding innovation, we analysethe way automatic learning breakthroughs and the ability of the GenAI to change data analysis and decision-making processes and interact with the machine. Real-world deployment is being applied in key areas such as personal health diagnosis, smart financial forecasts, dynamic commitments from retail customers, and smart production systems. However, these technological dance moves come with significant challenges, especially related to the transformation of the labourforce, the moral significance of automatic decision-making, data security issues and the need for an updated legal framework. Our analysis shows that solving these problems requires cooperation among decision-makers, business leaders and technology developers. By looking at future trends, we find exciting advancements in AI systems that are user-friendly, training models for a flexible workforce, and strategies for integrating AI that ensure innovation is balanced with social responsibility. This assessment, by synthesizing current experimental studies with real-world data, offers valuable insights for companies pursuing AI applications. Weemphasize the importance of developing AI solutions that focus on improving human abilities rather than replacing them, and we propose active management methods to achieve this goal. The results provide a roadmap for responsible AI implementation in American companies and the basis for future research in this rapidly developing field.

2025, Jurnal Komputer, Informasi dan Teknologi

This research explores the integration of custom Artificial Intelligence (AI) models in Content Management Systems (CMS) for content creation, curation, and management. The primary objective is to examine how AI-driven solutions, tailored to specific organizational needs, can optimize content workflows, improve productivity, and personalize content at scale. The study also investigates the ethical considerations, challenges, and potential benefits associated with the use of AI in CMS.The research adopts a mixed-methods approach, combining both quantitative and qualitative data. Quantitative data was gathered through surveys distributed to content creators and managers who have experience with AI tools, measuring productivity improvements, time savings, and user satisfaction. Qualitative data was collected through semistructured interviews, offering deeper insights into the integration process, human oversight, and ethical issues related to AI-generated content.Results show that custom AI models significantly enhance content production efficiency, with respondents reporting increased content output and substantial time savings. The integration of AI also led to higher user satisfaction, particularly due to the personalized and relevant content generated by AI tools. However, challenges such as data quality, model bias, and the need for continuous training were identified. Ethical concerns regarding AI-generated content, including potential biases and intellectual property issues, were also highlighted.The study concludes that AI models tailored to organizational needs provide substantial benefits in terms of scalability, personalization, and efficiency. However, businesses must address the ethical implications and ensure proper human oversight to mitigate biases and ensure content quality and responsibility. Future research should focus on refining AI model transparency and inclusivity.

2025, Journal of Informatics Education and Research

The main focus of this research is to explore the use of machine learning algorithms to carry out the predictive analytics and hence it serves as a hedging tool for leadership decision-making and strategy development. With decades of structured and unstructured data from organization, you can predict future trends in the market, operational risks as well as customer behavior in with higher confidence. Continuous learning of data patterns with Tree Decision as well as Neural Network predictions, also include Artificial Intelligence (AI) predictions. It provides strategic forward looking for executives to make tactical decision. The ability to include predictive analytics in leadership allows this to improve pro active development of strategy and efficient resource allocation as well as reduce uncertainty within today's changing business landscape. This one way enables one get a culture of facts on which strategic choices are no longer held on blind guesses but on facts. The research concludes that either to address the challenges and/or to capture opportunities which emerge across business domains, business leaders with machine learning supported predictive analytics tool set enables them to better respond to the challenges and exploit opportunities, thus deriving the sustained competitive advantage to drive the growth. Quantifiable results are demonstrated in the case of in-place studies related to increased strategic agility as well as long term performance.

2025, THE ASIAN BULLETIN OF BIG DATA MANAGMENT

In the era of smart city development, enhancing accuracy and operational efficiency at Realtime is crucial. To meet the complexity of advanced urban environments smart city technologies-including IoT and ICT-play a vital role. These innovations aim to improve quality of life and represent progress in urban monitoring and control. Key sectors such as traffic management, healthcare, surveillance, governance, and security are becoming increasingly intelligent. As urban populations grow, the demand for integrated and efficient smart city services-particularly in surveillance, transportation, public safety, and healthcare-continues to rise. This paper proposes a concept for comprehensive and centralized smart city model that unifies various autonomous departments to operate cohesively under an AI-driven system, thereby enhancing responsiveness and resource coordination.

2025, Madhya Bharti (मध्य भारती) - Humanities and Social Sciences UGC Care Group I Journal

Socioeconomic determinants play a significant role in predicting the leading public health issues in India, Child malnutrition. This research examines the impact of economic status, parental education, maternal health, and availability of resources on child weight, drawing evidence from the National Family Health Survey (NFHS-5) for Telangana. Statistical Analysis and Machine Learning models, map the relationship of independent variables like Type of Residence, Source of Drinking Water, Type of Toilet Facility, Ethnicity, and Wealth Index Deciles with Child Weight. Chi-square tests identify important associations between child weight and major socioeconomic determinants, such as wealth index (χ² = 20.123, p = 0.0000), type of toilet facility (χ² = 3975.842, p = 0.00000), and Source of Drinking Water (χ² = 4173.472, p = 0.00000). For forecasting the risk of malnutrition, a comprehensive analytical framework incorporates Machine Learning methodologies, including Linear Support Vector Machines (15.436 for 80:20 partition), multiple linear regression model (15.311 for the 80:20 partition), and Decision Tree Regression (15.234 for 80:20 partition), were utilized, with Random Forest (14.771 for 80:20 partition) registering the lowest RMSE. This data-informed research perspective offers important insights for policymakers and public health authorities to help and contribute to minimizing malnutrition and achieving improved health status for children in Telangana.

2025, European Journal of Computer Science and Information Technology

The convergence of artificial intelligence and cloud computing has fundamentally transformed enterprise digital operations, delivering unprecedented improvements in efficiency, security, and cost management. Our analysis reveals that organizations implementing AI-powered cloud automation have achieved remarkable results: up to 85% reduction in manual operations, 42% decrease in operational costs, and 56% improvement in service quality. Through extensive case studies across manufacturing, healthcare, and financial services sectors, we demonstrate how AI-driven solutions are revolutionizing critical operations including predictive maintenance, resource optimization, and security threat detection. The research indicates that organizations leveraging these technologies have experienced a 67% improvement in system failure prediction and 73% reduction in downtime, while achieving 95% accuracy in pattern recognition and anomaly detection. This paper examines the architectural frameworks, implementation strategies, and best practices that enable these transformative outcomes, providing a comprehensive roadmap for organizations seeking to harness the synergy between AI and cloud computing for operational excellence.

2025

This study investigates the primary risk factors for hypertension in Malawi using predictive analytics and the CARROT-BUS (Capacity Building, Accountability, Resources, Results, Ownership, Transparency – Bottom-Up Strategy) model as a... more

2025, World Journal of Advanced Engineering Technology and Sciences

Artificial intelligence is transforming the financial services industry through revolutionary applications in risk management and fraud detection. This transformation extends beyond incremental improvements to fundamentally reimagine core financial processes, enabling institutions to process vast quantities of data, identify complex patterns, and make decisions with unprecedented speed and accuracy. AI-driven systems have evolved risk assessment beyond traditional statistical models by analyzing billions of variables simultaneously and detecting subtle correlations invisible to human analysts. In fraud detection, sophisticated anomaly detection algorithms establish individualized behavioral baselines for each customer, dramatically reducing false positives while preserving legitimate transactions. These systems identify fraudulent patterns in real-time, detect novel schemes, and recognize coordinated fraud rings with remarkable precision, translating directly to significant reduction in fraud losses and increased transaction volumes. Behavioral analytics has created unparalleled visibility into customer financial patterns, supporting both enhanced fraud prevention and hyper-personalized service offerings. As these technologies continue to mature, financial institutions must balance innovation with ethical considerations and regulatory compliance, recognizing that trustworthiness represents a powerful competitive advantage in an increasingly algorithm-mediated landscape.

2025, Avances en Energías Renovables y Medio Ambiente - AVERMA

Los peces son afectados por las temperaturas bajas del agua durante los meses invernales, lo que ocasiona disminución de la tasa de reproducción y de la producción en la acuicultura. Se efectuaron ensayos en dos modelos de estanques, en el primero se colocaron materiales flotantes y el segundo funcionó como testigo. Buscando morigerar los descensos de temperatura del agua, se ensayaron botellas de PET y captadores solares flotantes con doble cubierta de acetato. Las experiencias fueron efectuadas bajo las condiciones climáticas del NEA, durante el invierno y la primavera. El captador solar flotante tuvo una respuesta aceptable respecto al incremento de temperatura buscado. Pueden obtenerse efectos mejores maximizando el área de la superficie transparente captadora de la radiación y disminuyendo el área de los bordes del captador. Se presentan las diferencias de temperatura entre las dos piletas de ensayo para los distintos modelos de captadores, además de otros parámetros climáticos. Palabras clave: Acuicultura, captadores solares flotantes, estanques piscicultura. La piscicultura experimenta actualmente un creciente desarrollo en el NEA, pero adolece aún de algunos problemas que dificultan un mayor crecimiento. Dentro de los problemas a resolver y que atentan contra el desarrollo de la actividad, se encuentra el estancamiento del crecimiento de los peces durante el invierno debido a las bajas temperaturas. Cuando la temperatura del agua desciende por debajo de los 20 °C, los peces dejan de alimentarse, acotando el período de cultivo a 200-250 días dependiendo de las condiciones locales. A ello debe sumarse que en situaciones de fríos extremos y prolongados, la temperatura del agua del estanque puede descender a menos de 10 °C, favoreciendo la aparición de enfermedades e inclusive ocasionando la muerte de los peces. Existen varios parámetros críticos en la acuicultura, ellos son la temperatura, el pH, el oxígeno disuelto y el amoníaco. Estos se tienen que medir diariamente o a lo largo de todo el día en el caso de sistemas intensivos de producción continua. Estos parámetros influencian las propiedades físicas y composición química del agua y consecuentemente un correcto manejo de los mismos puede mejorar el comportamiento de los peces (salud y crecimiento). Por el contrario, si no son correctamente atendidos, las consecuencias pueden ser serias, yendo desde bajas tasas de crecimiento hasta stress y mortalidad. El uso térmico de la energía solar para sistemas de calentamiento de agua es una de las tecnologías más difundidas a nivel mundial. Los invernaderos son los sistemas más aplicados ya que proveen incrementos significativos en la temperatura del agua, las mantas solares también presentan resultados similares. Sin embargo, hay poca información disponible sobre captadores solares probablemente debido a su elevado costo inicial. Los invernaderos constituyen la opción más difundida para ASADES

2025, Avances en Energías Renovables y Medio Ambiente

Las máquinas de absorción de doble etapa se pueden accionar a 2 temperaturas diferentes siendo apropiadas para un accionamiento mediante energía solar a baja temperatura (90ºC) y combustión de gas a alta temperatura (170ºC). Se presenta la descripción técnica de una enfriadora de agua por absorción de Agua-LiBr, de doble etapa, los resultados obtenidos mediante la modelización y los primeros resultados experimentales obtenidos en un banco de ensayo. En la modelización se utilizó el software EES en el que se cargaron los balances de materia, energía y las ecuaciones de transmisión de calor, utilizándose las propiedades de la mezcla definidas en el software. Se obtuvo la variación de la eficiencia en función del factor solar. La caracterización en un banco de ensayos permitió determinar que la eficiencia de la enfriadora accionada con gas natural es de 1.11 mientras que si es accionada en forma combinada con un factor solar de 0.26 la eficiencia es de 0.83.

2025, Proceedings of the 7th Unconventional Resources Technology Conference

Complexities in petrophysical and compositional properties as well as significant spatial heterogeneity of rock properties make formation evaluation challenging in organic-rich mudrocks. Conventional methods often overlook the importance of integrated rock classification for evaluation of formation properties, resulting in high uncertainties in estimates of mineralogy, porosity, fluid saturations, and total organic carbon content (TOC). The objectives of this paper include (a) developing an iterative workflow to simultaneously enhance formation evaluation and rock classification, (b) using the estimates of petrophysical, compositional, geochemical, and mechanical properties for completion-oriented rock classification to improve production decisions, and (c) using field-scale geostatistical analysis to extend the introduced workflow to neighboring wells without core measurements and minimizing model calibration efforts, while maintaining reliable formation evaluation results. First, we perform a joint inversion of well logs for depth-by-depth estimation of volumetric concentrations of minerals, porosity, TOC, mechanical properties, and fluid saturations by integrating information about thermal maturity and core/well-log measurements. These initial estimates are used for a preliminary petrophysical rock classification. Model parameters are updated in each rock class and are used in the second iteration for a class-by-class-based assessment of petrophysical, compositional, and mechanical properties. Spatial geostatistical analysis of formation properties is then used to select the range of neighboring wells where the developed models in each rock type is reliable. This iterative procedure is repeated until convergence of petrophysical/compositional properties in two subsequent iterations or agreement with core measurements (if available) is achieved. Finally, we perform an integrated completionoriented rock classification to determine the best rock types for completion. We successfully applied the proposed workflow to more than 100 wells in 20 different counties in the Midland Basin, 7 of which contained core and geochemical data. Results showed that the iterative workflow significantly improved estimates of TOC, porosity, and water saturation by approximately 56%, 28%, and 53% respectively, compared to a conventional method. Results also confirmed that the proposed workflow significantly enhances the formation evaluation and enables reliable reservoir characterization and URTeC 656 2 completion decisions in organic-rich mudrocks. Integrated geostatistical analysis, rock classification, and the advanced iterative formation evaluation workflow, is a novel approach which enables (a) reliable application of the developed rock physics models to wells with no core data or ECS logs and (b) incorporation of spatial heterogeneity of the formation for a reliable field-scale reservoir characterization.

2025, HAL (Le Centre pour la Communication Scientifique Directe)

Uber, Waze, Airbnb… The algorithms that control these platforms are based on an optimisation of the service provided to the user rather than any collective, political or moral norms. The accusations against these algorithms expose the way technical architectures implicitly govern our lives Journalists in England, Italy and France1 create false profiles for restaurants, and manage to push them up in Tripadvisor's rankings thanks to flattering comments and high marks. This is a way of denouncing the artificial popularity calculations that deceive clients and encourage unfair competition between restaurants. Researchers denounce the presence of 'phantom cars' on the Uber application,2 proving that the ride sharing company simulates the supply and demand market that it claims to display as an unbiased intermediary, in order to control rates and give users the impression of an abundant offer. These two accusations against digital calculations are typical of the initiatives seeking to audit and criticise digital platforms and their algorithms. They reveal the growing concern among public authorities, market actors and citizens regarding the space these platforms occupy in our daily lives. With the rampant spread of mobile phone equipment (4,8 billion in 2017, 2,32 billion of which 1 "Comment 'L'OEil du 20 heures' a piégé TripAdvisor en créant un faux restaurant", France Info, 07/09/15; "Un journal italien invente un faux restaurant pour piéger le site TripAdvisor", France Info, 05/07/15.

2025

The objective of this exploratory research paper is to investigate how organizations can utilize cloud computing models, Software-as-a-Service (SaaS), Platform-as-a-Service (PaaS), and Infrastructure-as-a-Service (IaaS), to scale up their growth and efficiency, lower costs and improve scalability. Using the Technology-Organization-Environment (TOE) framework, this research investigates the strategic advantages, adoption of drivers, and limitations of cloud computing. A qualitative research methodology was used, gathering primary data (100) from food service representatives via questionnaires and evaluating them using a structured framework. Results showed that cloud adoption lowered costs and had a positive environmental effect while improving operational efficiency, scalability, and mobility. Businesses gained from automatic backups and real-time data accessibility, which promote business continuity. However, issues with liability, compliance, and data security presented difficulties, particularly for smaller businesses. Large corporations fully integrated cloud computing (100%), whereas medium and small firms had lower acceptance rates (60% and 25%, respectively), according to a comparative examination of cloud adoption across all company sizes. Additionally, investment patterns indicated that larger businesses spent far more on cloud solutions. In this study, the research paper emphasizes how cloud computing promotes digital transformation, lowers carbon emissions, and maximizes the use of IT resources. Strategic adoption can improve business sustainability even while security threats and regulatory issues still exist. To optimize advantages, organizations must match their cloud plans with their operational objectives. To reduce adoption hurdles, future studies could examine sophisticated security systems and legislative initiatives.

2025, Journal of Posthumanism

This research examines how MIS frameworks strengthen energy infrastructure resilience through consolidated use of predictive models alongside data analytics and crisis management resources. There are multi factors, such as escalating natural disasters and elevated cyber threats with aging infrastructure systems, constantly push. The U.S. energy system toward declining resilience levels. The strategic decision-making and performance enhancement now depends heavily on Management Information Systems. This study uses qualitative research methods and relies on secondary data from energy reports alongside energy grid failure analysis and MIS implementation studies. This analysis reviews various MIS systems, such as SCADA and ERP, to identify how they could improve monitoring operations and evaluation procedures and quick response functionality. MIS development based on specific system needs leads to greater energy system surveillance capabilities and better resource management with improved recovery protocols. The research demonstrates how energy infrastructure protection improves when intelligent MIS combines real-time analysis along with predictive artificial intelligence technologies towards reaching national security goals for U.S. energy systems.

2025

In today's fast-moving, data-driven world, the future of Human Capital Management (HCM) hinges on intelligent, integrated, and predictive systems. At the center of this transformation is Business Intelligence (BI)-not just a dashboard or a report, but a framework for real-time, strategic decision-making. "Business intelligence is no longer about data collection. It's about insight-driven action."-Thomas H. Davenport, Author of Competing on Analytics 1 F 0 D What Is Business Intelligence in Human Capital?