NEWS.md
In sdim: An R Package for Supervised Dimension Reduction

sdim 0.1.0

Initial CRAN release.
Five factor extraction methods for asset pricing and macroeconomic forecasting:
- Principal Component Analysis (pca_est())
- Partial Least Squares (pls_est())
- Scaled PCA (spca_est()), Huang, Jiang, Li, Tong, and Zhou (2022)
- Reduced-Rank Approach (rra_est()), He, Huang, Li, and Zhou (2023)
- Instrumented PCA (ipca_est()), Kelly, Pruitt, and Su (2019)
Unified sdim_fit() wrapper with print(), summary(), plot(), and predict() methods.
Factor evaluation utility (eval_factors()) reporting R-squared, trace R-squared, and canonical correlations against benchmark factors.
Out-of-sample forecasting helpers: oos_standardize(), select_ar_lag_sic(), estimate_ar_res(), estimate_ardl_multi().
Bundled datasets replicating results from the source papers: grunfeld, he2023_factors, he2023_ff17vw, he2023_ff30vw, he2023_ff48vw, he2023_ff48ew, he2023_ff5, he2023_dacheng202, huang2022_ip, huang2022_macro.
Four vignettes: package overview, Huang et al. (2022) Table 4 replication, He et al. (2023) Table 3 replication, and IPCA on the Grunfeld panel.