发表机构
National Tsing Hua University; Stanford University（国立清华大学; 斯坦福大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究探索从投资者文本预测最大可能利润的可行性，提出三轨集成方法，发现波动率缩放与立场修正可提升预测，但需结合市场机制综合评估。
AI 中文摘要
BAIBAICHUCHU 团队参加了 NTCIR-19 FinArg-3 的社交媒体子任务，根据最大可能利润（MPP）对中国投资者帖子进行排序。一个由词汇特征、FinArg-2 预微调的 MacBERT 排序器和 LLM 评判器组成的三轨集成在帖子分组开发评估中达到 0.734，但我们最佳官方提交得分为 0.517。所有十二个提交的运行结果介于 0.4598 和 0.5402 之间，我们的 26 个三轨一致对得分为 0.500。事后审计发现，提交的评判器对看跌帖子应用了只做多规则，尽管 MPP 是立场感知的。纠正此问题改变了 87 个官方预测中的 28 个，但未改变准确率，却将开发准确率从 0.680 降至 0.622：一种特定于机制的语义捷径改善了验证拟合。我们将排序分解为方向性文本、波动率和时间跨度，以及成对边际。一个运行极值模型预测 $\sigma\sqrt{T}$ 缩放，这在来自 2026 年 7 月单独收集的 502 个价格对齐帖子中事后观察到。帖子发布前波动率按帖子特定观察时间跨度缩放，$\sigma_{\mathrm{pre},20}\sqrt{N_i}$，与后来的截断时间跨度 MPP 代理相关（Spearman $\rho=0.320$；ticker 聚类 95% CI [0.133,0.466]），并正确排序了 61.4% 的不平等结果对。转移文本与 7 月结果几乎不相关（$\rho=0.055$），并且在此得分和立场条件下增加的信息很少。开发可靠性随着标记的 MPP 差距扩大而从约 0.60 升至 0.93。早期 ERAI 结果 0.6207 排除了普遍不可预测性作为简单解释。评估应同时考虑文本、市场机制、历史波动率、时间跨度和对组成。
英文摘要
The BAIBAICHUCHU team participated in the Social Media Subtask of NTCIR-19 FinArg-3, ranking Chinese investor posts by Maximum Possible Profit (MPP). A three-track ensemble of lexical features, a FinArg-2-pre-finetuned MacBERT ranker, and an LLM judge reaches 0.734 in post-grouped development evaluation, but our best official run scores 0.517. All twelve submitted runs lie between 0.4598 and 0.5402, and our 26 unanimous three-track pairs score 0.500. A post-hoc audit finds that the submitted judge applied a long-only rule to bearish posts although MPP is stance-aware. Correcting it changes 28 of 87 official predictions without changing accuracy, yet lowers development accuracy from 0.680 to 0.622: a regime-specific semantic shortcut improved validation fit. We decompose ranking into directional text, volatility and horizon, and pairwise margin. A running-extremum model predicts $σ\sqrt{T}$ scaling, observed ex post in 502 price-aligned posts from a separate July 2026 collection. Pre-posting volatility scaled by the post-specific observation horizon, $σ_{\mathrm{pre},20}\sqrt{N_i}$, is associated with a later truncated-horizon MPP proxy (Spearman $ρ=0.320$; ticker-cluster 95% CI [0.133,0.466]) and correctly orders 61.4% of unequal-outcome pairs. Transferred text is nearly uncorrelated with the July outcome ($ρ=0.055$) and adds little conditional on this score and stance. Development reliability rises from about 0.60 to 0.93 as the labeled MPP gap widens. An earlier ERAI result of 0.6207 rules out universal unpredictability as a simple explanation. Evaluation should jointly consider text, market regime, historical volatility, horizon, and pair composition.
Comments15 pages, 1 figure, 6 tables. Both authors contributed equally. Selected for oral presentation at NTCIR-19 (FinArg-3 Social Media Subtask)
