FunnelCausalNet：面向多层优惠券分配的漏斗感知联合转化-收益 uplift 模型
FunnelCausalNet: Funnel-aware Joint Conversion-Revenue Uplift for Multi-tier Coupon Allocation
浏览论文内容
中文总结 AI 辅助
FunnelCausalNet 是一种耦合转化与价值头的 uplift 估计器，在半合成与工业数据集上均展现出优于基线的 GMV 效应误差降低及 DeltaROI 表现，为多层优惠券分配提供了漏斗感知的分配方案。
中文摘要 AI 辅助
优惠券活动旨在同时提升转化量与收益，但商品交易总额（GMV）遵循从转化到条件订单价值的确定性漏斗结构，且呈现零膨胀与重尾分布特性。本文提出 FunnelCausalNet，这是一种 uplift 估计器，通过 μ_gmv=μ_convμ_val 将二分类转化头与非负条件价值头耦合在一起。在明确的随机对照试验（RCT）、支持度、率差及跨头协方差控制假设下，理想化的主导阶均方误差（MSE）对比识别出一种机制，其中漏斗构成可降低逐点方差；这是一种启发式结论，而非共享表示神经网络模型的保证。该估计器搭配边际拆分共形 CATE 摘要，通过 Bonferroni 并集作为审计带，以及使用 RCT 锚定估计值的拉格朗日预算分配器，用于考虑补贴的投资回报率（ROI）核算。在半合成多层 Criteo-MT7 数据集上，FunnelCausalNet 的平均 AUUC_GMV 在 11 个基线模型中与领先的特征交互基线处于一个种子标准差范围内，而受控的 ablation 实验在测试的零膨胀机制下，相比直接 GMV 回归将 GMV 效应误差降低了 18%至 48%。在去标识化的工业酒店优惠券 RCT 日志（每个种子约 490 万保留曝光记录）上，预期结果评估遍历整个线性规划（LP）前沿；FunnelCausalNet 在 10%至 60%的 7 个相关锚点中均具有最佳的种子平均 DeltaROI，我们将此视为描述性前沿一致性而非独立显著性。在稀疏二元支出公共基准上，以收益为导向的排序器可优于 uplift 曲线代理，定义了明确的机制边界。
英文摘要
Coupon campaigns seek to lift both conversion and revenue, but gross merchandise value (GMV) follows a deterministic funnel from conversion to conditional order value and is zero-inflated and heavy-tailed. We propose FunnelCausalNet, an uplift estimator coupling a binary conversion head with a nonnegative conditional-value head through $μ_{\mathrm{gmv}}=μ_{\mathrm{conv}}μ_{\mathrm{val}}$. Under explicit RCT, support, rate-gap, and cross-head covariance-control assumptions, an idealized leading-order MSE comparison identifies a regime in which funnel composition can reduce pointwise variance; this is a heuristic, not a guarantee for the shared-representation neural model. The estimator is paired with marginal split-conformal CATE summaries, combined through a Bonferroni union as audit bands, and a Lagrangian budgeted allocator using RCT-anchored estimates for subsidy-aware ROI accounting. On semi-synthetic multi-tier Criteo-MT7, FunnelCausalNet's mean AUUC_GMV is within one seed standard deviation of the leading feature-interaction baseline among eleven baselines, while a controlled ablation reduces GMV effect error versus direct GMV regression by 18--48% across tested zero-inflation regimes. On de-identified industrial Hotel-Coupon RCT logs with about 4.9 million hold-out exposure records per seed, expected-outcome evaluation sweeps full LP frontiers; FunnelCausalNet has the best seed-averaged mean DeltaROI at all seven correlated anchors from 10% to 60%, which we treat as descriptive frontier consistency rather than independent significance. On sparse binary-spend public benchmarks, revenue-focused rankers can dominate uplift-curve proxies, defining an explicit regime boundary.
发表机构
- AMap Alibaba Group（阿里巴巴高德集团）
机构由 AI 辅助整理，请以论文原文为准。
