arXiv 2004.065652020-04q-fin.STcs.LGstat.APstat.ML

Bayesian Consensus：异方差噪声下误校准仪器的一致估计

Bayesian Consensus: Consensus Estimates from Miscalibrated Instruments under Heteroscedastic Noise

J.P. Morgan AI Research（摩根大通人工智能研究院）；Carnegie Mellon University（卡内基梅隆大学）

Chirag Nagpal, Robert E. Tillman, Prashant Reddy, Manuela Veloso

发表信息：: NeurIPS 2019 Workshop on Robust AI in Financial Services: Data, Fairness, Explainability, Trustworthiness and Privacy

摘要：

我们认为，从一组人类预测，模型，传感器或其他仪器，可能会受到偏见或误校准和随机异方差噪声的预测或测量聚合的问题。我们提出了一个贝叶斯一致性估计，调整误校准和噪声，并表明该估计是无偏的，渐近比天真的替代品更有效。我们进一步提出了一个层次贝叶斯模型，利用我们提出的估计，并将其应用于两个现实世界的预测挑战，需要共识估计从容易出错的个人估计：预测流感样疾病（ILI）每周的百分比和预测上市公司的年度收益。我们证明了我们的方法在减少偏差和错误方面是有效的，并且比现有的共识模型产生更准确的预测。

英文摘要：

We consider the problem of aggregating predictions or measurements from a set of human forecasters, models, sensors or other instruments which may be subject to bias or miscalibration and random heteroscedastic noise. We propose a Bayesian consensus estimator that adjusts for miscalibration and noise and show that this estimator is unbiased and asymptotically more efficient than naive alternatives. We further propose a Hierarchical Bayesian Model that leverages our proposed estimator and apply it to two real world forecasting challenges that require consensus estimates from error prone individual estimates: forecasting influenza like illness (ILI) weekly percentages and forecasting annual earnings of public companies. We demonstrate that our approach is effective at mitigating bias and error and results in more accurate forecasts than existing consensus models.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。