Package: cat2cat 0.6.1.9000

ML weights | Multi-period chaining | Optional: passing mappings$freqs_df | Mapping table stability and truncation | Backward chaining | Forward chaining | Adding ML to the chain | Panel data with subject identifiers | Aggregated data and special cases | Regression on replicated data | Building a 4-period harmonised repeated cross-section dataset | Neutral impact demonstration | Fixed effects regression | Choosing the right advanced workflow | Next steps

Step 1: Understand the competing weight assumptions | If ML fails on some rows: on_fail and fail_warn | Step 2: Check whether conclusions are sensitive to the weight choice | Compare weight methods on the same mapped data | Compare pruning strategies only after comparing full weights | Compare ensemble compositions when no single method dominates | Step 3: Validate whether ML actually improves on simpler baselines | What cat2cat_ml_run() is doing | Minimal validation workflow | Compare multiple ML methods in one run | Inspect per-group diagnostics when methods disagree | Decision rules for interpreting the output

What is cat2cat? | How to read the documentation | The Problem | The Solution | Value Added of cat2cat | Harmonisation Assumptions and Sensitivity | Three weighting schemes | When each assumption matters | Reducing assumption dependence | When cat2cat won't help | If you need something more advanced | Use Cases | Key Concepts | Direction: Backward vs Forward | ML weights in practice | Weights | Quick Example | Forward mapping example | Naive vs Frequency Weights | Value added: pooled regression across both periods | Diagnostic Plot | Mapping table | Common hierarchical classifications | Truncating mapping tables from hierarchical codes | Learn More