13. Demo: Building a Strategy

Cd13639 C1 L1 DEMO 6 V1

Understanding the Confusion Matrix and Strategy Enhancement

The sixth demo focuses on leveraging the confusion matrix to refine trading strategy performance, connecting it to several critical metrics:

  • Metrics Introduced:

    • Accuracy: Measures correct predictions.
    • Precision: Evaluates prediction accuracy when predicting a positive return.
    • Recall: Assesses the ability to detect all positive predictions.

  • Data Preparation:

    • Reload existing data, including historical prices and RSI indicators.
    • Establish benchmarks like S&P 500 with metrics like CAGR and Sharpe Ratio.

  • Confusion Matrix Integration:

    • Define binary variables (y_test and y_pred) to calculate confusion matrix components.
    • Introduce SK learn library tools to compute accuracy, precision, and recall.

  • Performance Assessment:

    • Initial accuracy at 48%, precision at 53%, with low recall.

  • Strategy Tweaking:

    • Lower the investment threshold to increase trade frequency.
    • New strategy shows improved CAGR of 15% and Sharpe Ratio of 0.85.
    • Achieves higher recall with lower precision.

  • Conclusion:

    • Adjustments between precision and recall can boost strategy performance, emphasizing the relation between frequent trading and profitability.

Next Concept