account identification

To address the limited interpretability of e-commerce customer churn models and their inadequate support for fine-grained retention decisions, this paper proposes a three-stage analytical framework integrating explainable AI (XAI), survival analysis, and RFM-based behavioral segmentation. Methodologically, it synergistically combines SHAP-based feature attribution, the Cox proportional hazards model, and dynamic RFM clustering to enable churn root-cause interpretation, optimal intervention timing prediction, and precise identification of high-risk customers. Its key contribution lies in the first systematic integration of explanation-driven attribution analysis, temporal risk modeling, and behaviorally heterogeneous segmentation—overcoming the opacity limitations of conventional black-box models. Empirical evaluation demonstrates that the framework significantly enhances both predictive transparency and intervention efficacy: customer retention improves by 12.7%. It thus delivers verifiable, actionable, and data-driven decision support for personalized retention strategies.

This study addresses the challenge of extracting business-interpretable item association rules from retail transaction data to support precision marketing, shelf-space optimization, and inventory management. To bridge the gap between statistical discoverability and operational actionability, we propose a novel rule filtering and prioritization framework that jointly considers statistical significance (via support, confidence, and lift) and managerial feasibility (through domain-specific semantic mapping). Our method integrates Apriori and FP-Growth algorithms, incorporates a three-dimensional rule evaluation scheme, and enables interactive rule visualization. Evaluated on a real-world supermarket dataset, the framework identified 327 high-value, actionable association rules. Deployment yielded an 18.6% increase in cross-buying rate and a 22.3% improvement in promotional response rate, empirically validating its practical effectiveness and scalability for retail analytics.

In e-commerce marketing, conventional intervention targeting suffers from low efficiency in identifying high-impact users and struggles to jointly satisfy operational constraints (e.g., budget, coverage) and maximize incremental value. Method: This paper proposes a precise intervention framework that integrates uplift modeling with explicit constraint optimization. It is the first to jointly formulate causal uplift models (e.g., X-Learner) and mixed-integer programming (MIP) to enable value-driven, fine-grained user segmentation under hard constraints—including budget limits and regulatory compliance—moving beyond coarse-grained, response-rate–based targeting. Contribution/Results: The framework ensures interpretability and production readiness. Evaluated on a live A/B test involving millions of users and deployed in a top-tier e-commerce platform’s production system, it achieves an 18.7% improvement in incremental conversion value over state-of-the-art methods while strictly adhering to budgetary and compliance requirements. It has been scaled to full production deployment.

研究通过分析和分类客户信件来识别寿险取消原因，以提高保险公司对客户行为的理解，采用数据获取与准备等方法。

Business professionals—non-technical domain experts—lack appropriate tools and methodologies for effective what-if analysis (WIA), hindering data-informed decision-making. Method: We conducted a two-phase mixed-methods user study—comprising contextual interviews and in-situ task-based evaluations—to systematically characterize their analytical behaviors for the first time. Contribution/Results: Based on empirical findings, we propose three domain-grounded design principles: business-contextual data preparation, risk-aware assessment, and domain-knowledge integration. We implemented and validated these principles in an interactive visual analytics prototype. The study identifies three critical support gaps, empirically confirms that six classes of what-if techniques significantly improve decision efficiency and confidence, and yields eight actionable design guidelines for commercial business intelligence systems. This work bridges a key theoretical and practical gap in WIA research concerning non-technical users.

This study addresses occupational burnout among Security Operations Center (SOC) practitioners, often stemming from misalignment between job demands and individual capabilities. Drawing on flow theory, the authors conduct an inductive content analysis of 106 global SOC job postings to systematically map the prevalence of certifications (e.g., CISSP), technical skills (e.g., Python, Splunk), and soft skills—particularly communication skills, mentioned in 50.9% of listings. The research reveals, for the first time, a structured pattern in the skill and certification requirements of SOC roles. These findings provide empirical grounding for achieving challenge–skill balance, refining recruitment practices, and guiding professional development. Furthermore, the study advances the discourse on flow-aligned person–job fit and sets the stage for future investigations into the impact of artificial intelligence on SOC workforce dynamics.

The objective of this paper is to understand what characteristics and features of clinical data influence physician's decision about ordering laboratory tests or prescribing medications the most. We conduct our analysis on data and decisions extracted from electronic health records of 4486 post-surgical cardiac patients. The summary statistics for 335 different lab order decisions and 407 medication decisions are reported. We show that in many cases, physician's lab-order and medication decisions can be well predicted from a small subset of all features.

This study addresses the limitations of traditional B2B customer segmentation approaches, such as the RFM model, which rely on singular metrics and struggle to capture the complexity and dynamics of business interactions. To overcome this, the authors propose a dynamic, multi-criteria segmentation framework that extends RFM by incorporating stability and growth dimensions. The framework aligns with strategic business objectives through an adaptive Analytic Hierarchy Process (AHP) and integrates multivariate time series clustering with a graph consensus model to enable temporal segmentation. Evaluated on data from over 3,000 manufacturing enterprises, the approach demonstrates strong temporal robustness and significantly enhances the precision of customer strategy formulation through preference-driven dynamic clustering.

This study addresses the challenge users face in constructing, refining, and validating personalized evaluation criteria when making decisions based on reviews, particularly due to the frequent oversight of infrequent yet critical details. To tackle this issue, the authors propose a visual analytics system that integrates large language models with the Analytic Hierarchy Process (AHP) to automatically generate an initial decision model from user reviews and support human-AI collaborative iterative refinement of criteria, weights, and evidence provenance. The system introduces a novel coverage gap detection mechanism to identify missing evaluation dimensions, incorporates AHP-based consistency constraints for interactive weight adjustment, and enhances decision transparency through multi-level scorecards and exportable reports. Experimental results demonstrate that the system significantly improves users’ control over their evaluation criteria and the overall trustworthiness of their decisions.

Estimating nonparametric ranking-based discrete choice models from transaction data—accounting for both single- and multiple-purchase behaviors—requires handling an exponential number of consumer types, leading to prohibitive computational complexity. This work proposes the first dynamic programming–based column generation framework that efficiently enumerates relevant consumer types in such models. The approach features a novel subproblem that generalizes the linear ordering problem and incorporates acceleration techniques to enhance optimization efficiency. The method accommodates various model extensions and demonstrates substantial improvements over existing approaches on both synthetic and real-world datasets, achieving significantly faster computation while maintaining high estimation accuracy. Furthermore, it exhibits strong performance in downstream assortment optimization tasks.