Exploration of CGAN Pix2Pix Model with Image Data for S&P Stock Price Prediction
Scholarly Review Journal · 2022
DOI: 10.70121/001c.121661 ↗Kurz gefasst
Fachartikel in der Zeitschrift Scholarly Review Journal, 2022. Verfasst von Cindy Su. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
With the incredibly volatile stock market in 2020, many professional investors won big, but almost all average citizen investors lacked the resources to participate, such as investment knowledge, time, attention, and buying capital,, leaving them at a great disadvantage. There have been AI and prediction algorithms that have been developed that are making “smart” investing available to more people, but the algorithms are still difficult to understand as there is a code language barrier. The algorithms are also difficult to trust as they are opaque in how they process strictly numerical data into predictions. Thus, this experiment tested to see if an image databased algorithm like the Pix2Pix Conditional Generative Adversarial Network (CGAN) would be able to predict the next week’s stock prices. By focusing on image data, it promotes more understanding of the algorithm as common investors can more easily see patterns through visual representations of stock prices. It also explores the Gramian Angular Summation/Difference Fields (GAF) image representation method that could encode more pertinent relationships between data points than a direct candlestick image. Experiments ended up showing that the model’s prediction GAF image for the next week’s prices had a 74% similarity rate with the GAF image of the next week’s actual prices, suggesting that the image-based model and GAFs could be considered for future stock prediction.
Zitieren
APA
Su, C. (2022). Exploration of CGAN Pix2Pix Model with Image Data for S&P Stock Price Prediction. Scholarly Review Journal. https://doi.org/10.70121/001c.121661
Vancouver
Su C. Exploration of CGAN Pix2Pix Model with Image Data for S&P Stock Price Prediction. Scholarly Review Journal. 2022. doi:10.70121/001c.121661
BibTeX
@article{Su2022,
author = {Su, Cindy},
title = {Exploration of CGAN Pix2Pix Model with Image Data for S&P Stock Price Prediction},
journal = {Scholarly Review Journal},
year = {2022},
doi = {10.70121/001c.121661}
}
EUVIMED