机器学习预测股市_机器学习在股市中的表现如何

机器学习预测股市

When it comes to using machine learning in the stock market, there are multiple approaches a trader can do to utilize ML models. From determining future risk to predicting stock prices, machine learning can be used for virtually any kind of financial modeling.

它涉及到在股市中使用机器学习W¯¯母鸡，有多种方式交易者可以做到利用ML车型。 从确定未来风险到预测股票价格，机器学习几乎可以用于任何类型的财务建模。

In our previous articles, we delved into the usage of two Time Series models: SARIMAX and Facebook Prophet. We utilized both of these models to forecast the potential, future prices of Bitcoin. Check out the following article if you are interested:

在之前的文章中，我们深入研究了两个时间序列模型的用法： SARIMAX和Facebook Prophet 。 我们利用这两种模型来预测比特币的潜在未来价格。 如果您有兴趣，请查看以下文章：

In another article, we used classification models to classify stocks based on their performance in quarterly reports. You can read about the entire process of how we engineered different features from these reports and trained our classification models to determine investment decisions in the article below:

在另一篇文章中，我们使用分类模型根据股票在季度报告中的表现对股票进行分类。 您可以在以下文章中了解有关我们如何设计这些功能的不同功能以及训练我们的分类模型以确定投资决策的整个过程：

These are just a couple examples of how we can utilize machine learning models for financial markets.

这些只是我们如何在金融市场中利用机器学习模型的几个例子。

In these articles, we just used models to predict future prices, however, that is just half the battle. The next step is how do we evaluate these models if they were to actually be used while trading? The answer to that question is called — Backtesting.

在这些文章中，我们仅使用模型来预测未来价格，但这只是成功的一半。 下一步是如何在交易中实际使用这些模型时对其进行评估？ 该问题的答案称为-回测。

回测 (Backtesting)

What is backtesting? Backtesting is the process of applying a trading strategy, predictive model, or analytical method to historical data to evaluate its accuracy and performance.

什么是回测？ 回测是将交易策略，预测模型或分析方法应用于历史数据以评估其准确性和性能的过程。

It is very important to note that backtesting does not 100% accurately, represent live-trading in the past. However, you should only use it to inform your decision to live-trade your strategy or not. Even then, it may be more practical to forward-test your strategy before trading with real money.

非常重要的一点是要注意，回溯测试不能100％准确地表示过去的实时交易。 但是，您仅应使用它来告知您决定是否进行实时交易的决策。 即使这样，在使用真金白银进行交易之前，对您的策略进行前瞻性测试可能更为实用。

When we create a machine learning model, we need to backtest the model in order to determine how well it would have potentially performed in the past by feeding it historical data. There are two approaches when it comes to backtesting — Event-Driven and Vectorized Backtesting.

当我们创建机器学习模型时，我们需要对模型进行回测，以便通过提供历史数据来确定其在过去的潜在性能。 回溯测试有两种方法- 事件驱动和向量化回测。

向量化回测 (Vectorized Backtesting)

Today, we will be using vectorized backtesting in order to evaluate the performance of our machine learning model. This approach allows us to quickly observe how our ML model might have performed in the past. If you would like to learn more about vectorized backtesting, then we suggest reading the following article by a machine learning researcher which contributed to the outcome of this current project:

今天，我们将使用矢量化回测，以评估我们的机器学习模型的性能。 这种方法使我们能够快速观察过去ML模型的性能。 如果您想了解有关向量化回测的更多信息，那么我们建议阅读机器学习研究人员的以下文章，该文章为当前项目的成果做出了贡献：

编码我们的机器学习模型 (Coding Our Machine Learning Model)

In order to evaluate the performance of a machine learning model, we’ll first have to construct it in Python. The model we will be using is the AutoARIMA time series model from the