Algorithmic trading and machine learning have similar challenges. For example, suppose you create a trading strategy with a few hyperparameters that proves to be profitable during backtesting. When it comes to deploying the strategy to production, one question remains: "Is this set of parameters optimal?

Hyperparams Optimize

Deep learning has become a promising way to model the complexity of stock movements. It enables us to capture non-linear movements, associate large data, and reduce noise without an assumption of a pre-specified underlying structure. However, it also presents the challenge of selecting numerous hyperparameters, which critically affects the performance of the resulting models.

Human problem

Most studies dealing with financial time series typically choose pre-specified hyperparameters and check the robustness of the model based on small changes in the parameters. This approach requires experts to put a lot of effort into tuning numerous parameters simultaneously, which often results in a suboptimal model

HPO help

Hyperparameter optimization (HPO) can be used to mitigate this problem by automatically searching for the most optimal hyperparameters in machine learning models, and has been widely used to quickly identify good configurations.

    Our proposals and solutions

    We can provide clients with a layer of HPO for their own trading strategy.

    • Bayesian Method
    • "Sharpe Ratio" repport
    • Optimize best RoI, Buy & Sell signal
    • Around 1000 Epoch authorized (randomize possible)
    • Best hyperparams for all indicators
    • Professional Consulting Services
    • Dev Support Help can be provide
    • client Service & Operations

    Compare our HPO plans

    Click here and choose a plan that fits you.