Prediction of Stocks Index Price using Quantum GANs

TL;DR

This work investigates Quantum Generative Adversarial Networks (QGANs) for stock index prediction, comparing classical GANs, Hybrid Quantum GANs, and Fully Quantum GANs on FTSE-like data. It introduces a Fully Quantum GAN with a SWAP-test discriminator and an Invertible variant to address normalization challenges, and examines architecture choices such as rotation-angle encoding and amplitude embedding. Empirical results show that quantum approaches can achieve competitive or superior accuracy under certain conditions, with the Invertible FQGAN offering robust prediction in scenarios with limited data and resource constraints. The study demonstrates a concrete step toward quantum-accelerated financial forecasting, with implications for rapid, precise market analysis and portfolio decision-making, while also outlining current hardware and training limitations that guide future work.

Abstract

This paper investigates the application of Quantum Generative Adversarial Networks (QGANs) for stock price prediction. Financial markets are inherently complex, marked by high volatility and intricate patterns that traditional models often fail to capture. QGANs, leveraging the power of quantum computing, offer a novel approach by combining the strengths of generative models with quantum machine learning techniques. We implement a QGAN model tailored for stock price prediction and evaluate its performance using historical stock market data. Our results demonstrate that QGANs can generate synthetic data closely resembling actual market behavior, leading to enhanced prediction accuracy. The experiment was conducted using the Stocks index price data and the AWS Braket SV1 simulator for training the QGAN circuits. The quantum-enhanced model outperforms classical Long Short-Term Memory (LSTM) and GAN models in terms of convergence speed and prediction accuracy. This research represents a key step toward integrating quantum computing in financial forecasting, offering potential advantages in speed and precision over traditional methods. The findings suggest important implications for traders, financial analysts, and researchers seeking advanced tools for market analysis.

Figures (15)

• Figure 1: GAN Architecture from paper 29
• Figure 2: QGAN Architecture from paper 28
• Figure 3: Proposed QGAN Model for Stock Price Prediction
• Figure 4: Quantum circuit as a generator for Hybrid QGAN for a past window of 3 days and future of 1 day. Pauli-Z observable on the first qubit.
• Figure 5: Data preparation for Invertible FQGAN. We use overlapped train and test data to train the model(left). Once trained, the Quantum Generator can generate 16 values in the future, which have an overlap with the input data and therefore the normalization factor can be obtained (right).