Cross-Basis Construction

What problem does it solve?

This Skill simplifies the complex process of constructing bi-dimensional cross-basis matrices, which are essential for modeling both exposure-response and lag-response relationships simultaneously in Distributed Lag Non-Linear Models (DLNMs).

Core Features & Use Cases

• Flexible Basis Functions: Supports various basis functions like natural cubic splines (ns), B-splines (bs), and penalized splines (ps) for both exposure and lag dimensions.
• Optimized Knot Placement: Provides guidance on placing knots effectively for both exposure (e.g., percentiles) and lag (e.g., log scale) dimensions to avoid overfitting.
• Key Argument Guidance: Explains crucial arguments like lag, argvar, and arglag for accurate model specification.
• Use Case: When analyzing the impact of daily temperature on mortality, use this Skill to define how temperature variations and their effects over subsequent days are modeled.

Quick Start

Construct a cross-basis matrix for a daily temperature exposure with a 21-day lag, using natural cubic splines for both dimensions with appropriate knot placements.