Stock Volume Forecasting with Advanced Information by Conditional Variational Auto-Encoder

TL;DR

This paper introduces a Conditional Variational Auto-Encoder (CVAE) model for stock volume forecasting that leverages advanced input information to produce more accurate and scenario-rich non-linear time series predictions.

Contribution

The paper presents a novel application of CVAE for stock volume forecasting, incorporating advanced information like rebalancing dates to enhance prediction accuracy.

Findings

CVAE outperforms traditional linear models in accuracy.

Generated forecasts closely match actual data correlations.

Scenario generation aids in better interpretation of stock volume dynamics.

Abstract

We demonstrate the use of Conditional Variational Encoder (CVAE) to improve the forecasts of daily stock volume time series in both short and long term forecasting tasks, with the use of advanced information of input variables such as rebalancing dates. CVAE generates non-linear time series as out-of-sample forecasts, which have better accuracy and closer fit of correlation to the actual data, compared to traditional linear models. These generative forecasts can also be used for scenario generation, which aids interpretation. We further discuss correlations in non-stationary time series and other potential extensions from the CVAE forecasts.