Abstract:Financial time series prediction, whether for classification or regression, has been a heated research topic over the last decade. While traditional machine learning algorithms have experienced mediocre results, deep learning has largely contributed to the elevation of the prediction performance. Currently, the most up-to-date review of advanced machine learning techniques for financial time series prediction is still lacking, making it challenging for finance domain experts and relevant practitioners to determine which model potentially performs better, what techniques and components are involved, and how the model can be designed and implemented. This review article provides an overview of techniques, components and frameworks for financial time series prediction, with an emphasis on state-of-the-art deep learning models in the literature from 2015 to 2023, including standalone models like convolutional neural networks (CNN) that are capable of extracting spatial dependencies within data, and long short-term memory (LSTM) that is designed for handling temporal dependencies; and hybrid models integrating CNN, LSTM, attention mechanism (AM) and other techniques. For illustration and comparison purposes, models proposed in recent studies are mapped to relevant elements of a generalized framework comprised of input, output, feature extraction, prediction, and related processes. Among the state-of-the-art models, hybrid models like CNN-LSTM and CNN-LSTM-AM in general have been reported superior in performance to stand-alone models like the CNN-only model. Some remaining challenges have been discussed, including non-friendliness for finance domain experts, delayed prediction, domain knowledge negligence, lack of standards, and inability of real-time and high-frequency predictions. The principal contributions of this paper are to provide a one-stop guide for both academia and industry to review, compare and summarize technologies and recent advances in this area, to facilitate smooth and informed implementation, and to highlight future research directions.

Wasserstein distance loss function for financial time series deep learning

Deep Learning for Knowledge-Driven Ontology Stream Prediction.

Fine-Tuning a Time Series Foundation Model with Wasserstein Loss

Wasserstein GAN: Deep Generation applied on Bitcoins financial time series

Financial Time Series Prediction Using Deep Learning

Wasserstein autoregressive models for density time series

Deep Learning for Financial Time Series Prediction: A State-of-the-Art Review of Standalone and Hybrid Models

Financial Time Series Prediction Based on Deep Learning

GDTW: A NOVEL DIFFERENTIABLE DTW LOSS FOR TIME SERIES TASKS

A deep learning framework for financial time series using stacked autoencoders and long-short term memory

Wasserstein Adversarial Examples on Univariant Time Series Data

Learning Deep Time-index Models for Time Series Forecasting

WEITS: A Wavelet-enhanced residual framework for interpretable time series forecasting

Deep Video Prediction for Time Series Forecasting

Image Processing Tools for Financial Time Series Classification

AdaWaveNet: Adaptive Wavelet Network for Time Series Analysis

Entropic Dynamic Time Warping Kernels for Co-evolving Financial Time Series Analysis

Wavelet Neural Network Model with Time-Frequency Analysis for Accurate Share Prices Prediction

A seasonal-trend decomposition-based dendritic neuron model for financial time series prediction

Science Time Series: Deep Learning in Hydrology

Temporal Logistic Neural Bag-of-Features for Financial Time series Forecasting leveraging Limit Order Book Data