ML Training in Research Environment and Inference in Algorithms

Yes, it is possible to train a machine learning model in the research environment, store the trained models, and load them for prediction in the algorithm environment in QuantConnect.

After you train a model, you can save it into the Object Store. In QuantConnect Cloud, your Object Store data is backed up on QuantConnect servers. If you save models in live algorithms, save them at the end of the training method so you can access the trained model again if your algorithm stops executing. If you save models in backtests, save them during the OnEndOfAlgorithm event handler so that saving multiple times doesn't slow down your backtest.

You can load machine learning models from the Object Store or a custom data file like pickle. If you load models from the Object Store, before you load the model into your algorithm, in the Initialize method, check if the Object Store already contains the model.

If you train a model in the Research Environment, you can also save it into the Object Store to transfer it to the backtesting and live trading environment.

For sklearn models, you can save/read any data including sklearn models into/from objectstore by using JSON format. You will need a dictionary to save 2 key-value pairs within it:

• 1 is the model's initial parameters ```
• yourTrainedModel.get_params(), ``` For more information, you can refer to this docs and this stack overflow's thread.

Please note that to avoid look-ahead bias in backtests, don't train your model on the same data you use to test the model.