{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,10]],"date-time":"2026-10-10T07:49:03Z","timestamp":1791618543250,"version":"4.3.4"},"posted":{"date-parts":[[2026]]},"group-title":"SSRN","reference-count":0,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"abstract":"<jats:p>While the behavior of the stock market can be very unpredictable, the regime of the market is a major factor in its behavior. Being able to predict it can significantly aid in making profitable trading decisions in the stock market. This project investigated whether a machine learning model using daily price and volume data could predict next-day bull and bear regimes in the U.S. stock market in a way that produced trading returns exceeding a buy-and-hold strategy in the S&amp;amp;P 500. It tested a logistic regression model and a random forest model on a dataset of 6,339 daily samples of historical data from SPY, the exchange-traded fund that tracks the S&amp;amp;P 500, involving various price and volume features. The best total return achieved by a machine learning model was from the tuned logistic regression model, with 68.52%; however, the buy-and-hold strategy achieved a total return of 87.55%. While the model was outperformed by the buy-and-hold strategy in terms of total return, it did achieve its returns with less risk. The Sharpe ratio measures returns while also adjusting for risk. The buy-and-hold strategy had a Sharpe ratio of 0.82, but the tuned logistic regression model had a Sharpe ratio of 0.98. These results likely occurred because SPY generally trended upward during the testing period, so the machine learning strategies occasionally exited the market during predicted bear regimes and missed gains that the buy-andhold benchmark benefited from. The results suggest that machine learning may be more useful for risk management rather than consistently outperforming a buy-and-hold strategy in total return. However, future research could test the models on more volatile assets, utilize a higher number of technical indicators involving price and volume, and test additional machine learning models in order to determine whether stronger predictive performance can translate into higher trading returns when compared to a buy-and-hold benchmark.<\/jats:p>","DOI":"10.2139\/ssrn.7589460","type":"posted-content","created":{"date-parts":[[2026,10,10]],"date-time":"2026-10-10T07:04:53Z","timestamp":1791615893000},"source":"Crossref","is-referenced-by-count":0,"title":["Predicting Bull and Bear Market Regimes using Machine Learning:&amp;nbsp;A Supervised Classification Approach to U.S. Stock Market Trading"],"prefix":"10.2139","author":[{"ORCID":"https:\/\/orcid.org\/0009-0008-0971-388X","authenticated-orcid":true,"given":"Michael","family":"Thomas","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","container-title":[],"original-title":[],"deposited":{"date-parts":[[2026,10,10]],"date-time":"2026-10-10T07:04:53Z","timestamp":1791615893000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ssrn.com\/abstract=7589460"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":0,"URL":"https:\/\/doi.org\/10.2139\/ssrn.7589460","relation":{},"subject":[],"published":{"date-parts":[[2026]]},"subtype":"preprint"}}