Our autoencoder asset pricing model delivers out-of-sample pricing errors that are far smaller (and generally insignificant) compared to other leading factor models.doi:10.1016/j.jeconom.2020.07.009Shihao GuBryan KellyDacheng XiuJournal of Econometrics...
0.文章信息 文章标题:基于自编码器(Autoencoder)的资产定价模型(Autoencoder Asset Pricing Models) Shihao Gu, Bryan Kelly, Dacheng Xiu, Autoencoder asset pricing models, Journal of Econometrics, Vol…
见《Autoencoder asset pricing models》,使得可以处理不同输入尺寸的资产收益率向量,通过在一个standard的encoder-decoder架构中引入先验信息即截面资产的初始因子暴露度矩阵来帮助重建截面的资产收益率向量。
We propose a new latent factor conditional asset pricing model, which delivers out-of-sample pricing errors that are far smaller (and generally insignificant) compared to other leading factor models.
In addition to denoising, HOLO also applies Dropout during the training process. Dropout is a commonly used regularization technique primarily aimed at reducing model overfitting. In deep learning models, overfitting refers to the phenomenon where a model performs well on training data bu...
We propose a new latent factor conditional asset pricing model. Like Kelly, Pruitt, and Su (KPS, 2019), our model allows for latent factors and factor exposuresdoi:10.2139/ssrn.3335536Gu, ShihaoKelly, Bryan T.Xiu, DachengSocial Science Electronic Publishing...