Popular Libraries

Import the tslearn libraries.

from AlgorithmImports import *
from tslearn.barycenters import softdtw_barycenter
from tslearn.clustering import TimeSeriesKMeans

In the Initializeinitialize method, subscribe to some data so you can train the tslearn model.

# Subscribe to security data and store symbols for referencing in the algorithm.
tickers = ["SPY", "QQQ", "DIA", 
           "AAPL", "MSFT", "TSLA", 
           "IEF", "TLT", "SHV", "SHY", 
           "GLD", "IAU", "SLV", 
           "USO", "XLE", "XOM"]
symbols = [self.add_equity(ticker, Resolution.DAILY).symbol for ticker in tickers]

In this example, train a model that clusters the universe of Equities into distinct groups and then allocate an equal portion of the portfolio to each cluster. To cluster the securities, instead of using a real-time comparison, apply Dynamic Time Wrapping Barycenter Averaging (DBA) to their historical prices and then run a k-means clustering algorithm. DBA is a technique of averaging a few time-series into a single one without losing much of their information. Since not all time-series move efficiently like in ideal EMH assumption, this technique allows similarity analysis of different time-series with sticky lags. The following image shows a visualization of the process. For more information about the technical details, see Dynamic Time Warping in the tslearn documentation.

To perform DBA and then cluster the securities by k-means, create a TimeSeriesKMeans model:

# Cluster time series into 6 groups using DTW for similarity measurement.
self.model = TimeSeriesKMeans(n_clusters=6,   # We have 6 main groups
                    metric="dtw")

You can train the model at the beginning of your algorithm and you can periodically re-train it as the algorithm executes.

Warm Up Training Data

You need historical data to initially train the model at the start of your algorithm. To get the initial training data, in the Initializeinitialize method, make a history request.

# Initialize training data with a rolling window of size 252 days.
training_length = 252
self.training_data = {}
history = self.history(self.symbols, training_length, Resolution.DAILY).unstack(0).close
for symbol in self.symbols:
    self.training_data[symbol] = RollingWindow[float](training_length)
    for close_price in history[symbol]:
        self.training_data[symbol].add(close_price)

Define a Training Method

To train the model, define a method that fits the model with the training data.

# Prepare feature and label data for training by processing rolling window data to create time-series sequences for model training.
def get_features(self):
    close_price = pd.DataFrame({symbol: list(data)[::-1] for symbol, data in self.training_data.items()})
    log_price = np.log(close_price)
    log_normal_price = (log_price - log_price.mean()) / log_price.std()

    return log_normal_price

def my_training_method(self):
    features = self.get_features()
    self.model.fit(features.T.values)

Set Training Schedule

To train the model at the beginning of your algorithm, in the Initializeinitialize method, call the Traintrain method.

# Train the model initially to provide a baseline for prediction and decision-making.
self.train(self.my_training_method)

To periodically re-train the model as your algorithm executes, in the Initializeinitialize method, call the Traintrain method as a Scheduled Event.

# Train the model every Sunday at 8:00 AM
self.train(self.date_rules.every(DayOfWeek.SUNDAY), self.time_rules.at(8, 0), self.my_training_method)

Update Training Data

To update the training data as the algorithm executes, in the OnDataon_data method, add the current TradeBar to the RollingWindow that holds the training data.

# Add the latest bar to training data to ensure the model is trained with the most recent market data.
def on_data(self, slice: Slice) -> None:
    for kvp in slice.bars:
        self.training_data[kvp.key].add(kvp.value.CLOSE)

To predict the labels of new data, in the OnDataon_data method, get the most recent set of features and then call the predict method.

# Generate feature set and predict with the latest data for current market decisions.
features = self.get_features()
self.labels = self.model.predict(features.T.values)

You can use the label prediction to place orders.

# Use label prediction to place orders based on forecasted market direction.
for i in set(self.labels):
    assets_in_cluster = features.columns[[n for n, k in enumerate(self.labels) if k == i]]
    size = 1/6/len(assets_in_cluster)
    self.set_holdings([PortfolioTarget(symbol, size) for symbol in assets_in_cluster])

Follow these steps to save tslearn models into the Object Store:

1. Set the key name of the model to be stored in the Object Store.

# Set the key to store the model in Object Store for reuse across sessions.
model_key = "model.hdf5"

2. Call the GetFilePathget_file_path method with the key.

# Get the file path to correctly save and access the model in Object Store.
file_name = self.object_store.get_file_path(model_key)

This method returns the file path where the model will be stored.

3. Delete the current file to avoid a FileExistsError error when you save the model.

import os
os.remove(file_name)

4. Call the to_hdf5 method with the file path.

# Serialize Python objects into a file to save the model's state for other runs.
self.model.to_hdf5(file_name)

You can load and trade with pre-trained tslearn models that saved in Object Store. To load a tslearn model from the Object Store, in the Initializeinitialize method, get the file path to the saved model and then call the from_hdf5 method.

# Load the tslearn model from Object Store to use its saved state and update with new data if needed.
def initialize(self) -> None:
    if self.object_store.contains_key(model_key):
        file_name = self.object_store.get_file_path(model_key)
        self.model = TimeSeriesKMeans.from_hdf5(file_name + ".hdf5")

The ContainsKeycontains_key method returns a boolean that represents if the model_key is in the Object Store. If the Object Store does not contain the model_key, save the model using the model_key before you proceed.