Diffusion Variational Autoencoder for Tackling Stochasticity in Multi-Step Regression Stock Price Prediction

By Kelvin J.L. Koa et al

Published on Oct. 21, 2023

Read the original document by opening this link in a new tab.

Table of Contents

1. Introduction
2. Related Works
3. Methodology
3.1 Problem Formulation
3.2 Deep Hierarchical VAE
3.3 Input Sequence Diffusion
3.4 Target Sequence Diffusion

Summary

The document discusses the challenges of multi-step stock price prediction and introduces the Diffusion Variational Autoencoder (D-Va) model to address these challenges. The model combines a deep hierarchical variational-autoencoder and diffusion probabilistic techniques to predict stock prices through a stochastic generative process. By gradually adding noise to input and target sequences, the model aims to improve generalizability and prediction accuracy. Extensive experiments demonstrate the superiority of the D-Va model in prediction accuracy and variance over existing methods.