Statistical Learning Concepts for Quantitative Finance
Summary
This guide introduces statistical learning through an example of predicting the S&P 500 from company fundamentals. It frames the task as estimating a response from predictor features plus residual error, then distinguishes prediction, which prioritizes accurate outputs, from inference, which seeks to explain relationships and identify relevant predictors. It notes the tradeoff between interpretable linear models and potentially more predictive nonlinear approaches.
The guide contrasts parametric methods, which assume a functional form and estimate its parameters, with more flexible nonparametric methods that need more observations and can overfit. It also explains supervised learning with labeled outcomes and unsupervised learning for structure discovery such as clustering or dimensionality reduction. These are conceptual explanations rather than empirical comparisons: no model results or implementation details are supplied. The guide cautions that financial data’s low signal-to-noise ratio makes excess flexibility risky, even when historical datasets are large.
Key ideas
- Statistical learning estimates a relationship between predictor features and a response while accounting for residual error.
- Prediction focuses on accurate outputs, while inference focuses on explaining predictor-response relationships.
- Parametric models make functional assumptions; nonparametric models are more flexible but generally need more data.
- Both model families can overfit, a particular concern with noisy financial time series.
- Supervised learning uses labeled outcomes, while unsupervised learning seeks structure without labels.
Tags
Full text
# OctoBot 1.0.4 - What's new --- title: "OctoBot 1.0.4 - What's new" slug: "octobot-1-0-4-whats-new" date: "2023-12-10" authors: ["guillaume"] tags: ["OctoBot cloud", "Release", "DCA", "Exchanges"] image: "/images/blog/octobot-1-0-4-whats-new/with-octobot-1.0.4-use-octobot-cloud-strategies-and-trade-on-bingx.png" --- # OctoBot 1.0.4 - What's new  ## Introducing OctoBot 1.0.4 We are glad to announce the new release of OctoBot. 1.0.4 is a updated version adding the very anticipated download of OctoBot cloud strategies directly in your OctoBot, the addition of the <a href="https://bingx.com/en-us/invite/Z4UUVX/" rel="nofollow">BingX exchange</a> to the [OctoBot partner exchanges](/guides/exchanges#partner-exchanges---support-octobot) and many improvements. ## Downloading OctoBot cloud strategies Starting from OctoBot 1.0.4, you can now profit from <a href="https://www.octobot.cloud/explore" rel="nofollow">OctoBot cloud strategies</a> directly from your [OctoBot trading bots](https://www.octobot.cloud/trading-bot).  Directly from your OctoBot, download OctoBot cloud strategies and: - Use them with simulated or real funds - Configure them to trade differently, on other exchanges or other pairs - Backtest them using the OctoBot [backtesting engine](/guides/octobot-usage/backtesting) or the [Strategy Designer](/guides/octobot-usage/strategy-designer) available on [OctoBot trading bots](https://www.octobot.cloud/trading-bot) to optimize them according to your ideas ## BingX is now available on OctoBot At OctoBot, we work on making trading as accessible as possible. This comes with the support of most major exchanges. Following this philosophy, we just added support for <a href="https://bingx.com/en-us/invite/Z4UUVX/" rel="nofollow">BingX</a>. We hope that this addition will help many of our users. ## Exchanges bugfixes In OctoBot 1.0.4, fixed many exchange related issues, especially regarding futures trading and take profit / stop loss orders. Special thanks to Nes, Grr, Gerhard and Artem from our community who helped us a lot finding those issues. ## Other improvements and bugfixes In this release, we also added parameters to make the DCA and Daily trading mode more customizable and trade closer to your ideas. Many bugs have been fixed as well, especially regarding the web interface currency selector update, exchange connection issues, Ngrok configuration and more. ## Conclusion We can't wait to know what you think about this new version. Please use this <a href="https://feedback.octobot.cloud/open-source" rel="nofollow">feedback link</a> to share your suggestions and what you'd like to see in our next release.
Shown in full with attribution under the source's licence. Licence: GPL-3.0
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.