inflationkit 0.1.1

Patch release. One fix, to a side effect rather than to any returned value.

ik_sample_data() called set.seed(42L) and did not restore the previous state, so calling it reseeded the caller’s random stream: every later random draw in the session was silently reset to a fixed sequence. The package’s own test for this was failing. The seed now applies for the duration of the call only, and .Random.seed is restored on exit (or removed, if the session had none). The generated data are unchanged.

inflationkit 0.1.0

• Initial release.
• CPI component decomposition into weighted contributions via ik_decompose().
• Core inflation measures (trimmed mean, weighted median, exclusion-based, asymmetric trim) via ik_core(), following Bryan and Cecchetti (1994).
• Sticky vs flexible price decomposition via ik_sticky_flexible(), based on Atlanta Fed methodology (Bils and Klenow, 2004).
• Inflation persistence estimation (sum of AR coefficients, half-life via companion matrix, largest root) via ik_persistence().
• Diffusion indices via ik_diffusion().
• Phillips curve estimation (traditional, expectations-augmented, hybrid) via ik_phillips(). Newey-West HAC standard errors via robust_se = "HAC".
• Breakeven inflation from nominal and real yields via ik_breakeven().
• Trend inflation extraction (Hodrick-Prescott filter with frequency-based lambda, Beveridge-Nelson decomposition, exponential smoothing, moving average) via ik_trend().
• Forecast evaluation (Mincer-Zarnowitz bias, Nordhaus efficiency, Diebold-Mariano test) via ik_forecast_eval().
• Side-by-side comparison of core inflation measures via ik_compare().
• Built-in sample data via ik_sample_data().