报告详情
ECF-Net for Data Science and Intelligent Analytics of Employee, Customer and Financial Performance
编号：22
访问权限：仅限参会人
更新：2026-11-10 16:12:15 浏览：9次
Online
摘要
Employee performance, customer behaviour and financial performance jointly determine organisational success, yet analytics systems still model them in isolation and cannot say how much each domain contributes or how far a prediction can be trusted. This paper presents ECF-Net, a data-driven intelligent analytics framework that fuses human-resource, customer and financial indicators into a single multi-domain feature space, learns their nonlinear interactions with a deep neural network (DNN), and adds two new layers for trustworthy decision support: (i) a Triadic Synergy Index (TSI), derived from an exact Shapley–Harsanyi decomposition over the three domains, that quantifies value created only when all domains are modelled together; and (ii) a conformal decision layer that turns class probabilities into prediction sets with a distribution-free coverage guarantee. On a survey of 600 respondents, ECF-Net attains an accuracy of 98.42%, a ROC-AUC of 0.9951, an MCC of 0.9754 and an R² of 0.9912 with 2.6 ms inference per sample. On a controlled five-fold benchmark it is statistically on par with or better than four classical baselines; full fusion raises macro-F1 by 8.1 points over the best single domain; and TSI reveals that pairwise domain combinations are largely redundant (negative dividends) while the three-way interaction contributes +0.167 macro-F1. The conformal layer achieves 90.7% coverage at a 90% target with 1.43 classes per set, automating 57% of decisions at 85.7% accuracy and escalating the rest. ECF-Net thus delivers accurate, explainable and risk-aware organisational analytics that can be deployed on cloud or edge platforms.
关键词
Data-driven intelligent analytics; deep neural network; multi-domain feature fusion; Shapley attribution; triadic synergy; conformal prediction; trustworthy decision support
报告人
Subir Gupta
Sanaka Educational T Dept. of C.S.E稿件作者
全部评论
重要日期
-
会议日期
11月20日
2026
至11月22日
2026
-
10月31日 2026
初稿截稿日期
-
11月07日 2026
提前注册日期
-
11月22日 2026
注册截止日期
主办单位
United Societies of Science
承办单位
Pondicherry University
协办单位
IEEE Section
IEEE Madras Section
IEEE Madras Section
联系方式
- ICon INDIA 2026
- ic******@usssociety.org
发表评论