The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.
Author: Gabriel
Cabrera
License: MIT + file LICENSE
sdim is an R package for supervised dimension
reduction for asset pricing and macro forecasting. It implements five
factor extraction methods. PCA, PLS, and RRA take a multivariate target
(T×N returns matrix) and a matrix of factor proxies. sPCA takes a
univariate target and scales each proxy by its OLS slope on the target
before extracting principal components. IPCA extracts latent factors
from panel data using time-varying characteristics as instruments,
estimated via alternating least squares (ALS). Performance of extracted
factors can be evaluated with eval_factors(). The package
ships with seven he2023_* datasets (factor proxies and
portfolio returns) from the He et al. (2023) replication package.
# Install from GitHub (not yet on CRAN)
# install.packages("devtools")
devtools::install_github("gabbocg/sdim")| Function | Method | Reference |
|---|---|---|
pca_est() |
Principal Component Analysis (PCA) | He et al. (2023) |
pls_est() |
Partial Least Squares (PLS) | He et al. (2023) |
rra_est() |
Reduced-Rank Approach (RRA) | He et al. (2023) |
spca_est() |
Scaled PCA (sPCA) | Huang et al. (2022) |
ipca_est() |
Instrumented PCA (IPCA) | Kelly, Pruitt & Su (2019) |
library(sdim)
set.seed(42)
X <- matrix(rnorm(200 * 20), 200, 20) # T x L factor proxies
ret <- matrix(rnorm(200 * 30) / 100, 200, 30) # T x N returns (target)
# Fit each method
fit_pca <- pca_est(target = ret, X = X, nfac = 3)
fit_pls <- pls_est(target = ret, X = X, nfac = 3)
fit_rra <- rra_est(target = ret, X = X, nfac = 3)
print(fit_rra)
#> <sdim_fit [rra]>
#> Observations : 200
#> Predictors : 20
#> Factors : 3
# Evaluate factor quality (RMSPE and total adj-R² from He et al. 2023, §2.4)
eval_factors(ret = ret, factors = fit_rra$factors)
#> Factor Evaluation
#> ----------------------------------------
#> Portfolios 30
#> Factors 3
#>
#> Performance (He et al., 2023, §2.4)
#> ----------------------------------------
#> RMSPE 0.9875 (%)
#> Total adj-R² 2.9593 (%)
#> SR 0.0522
#> A2R 0.9443# Simulate panel
set.seed(99)
TT <- 120
K <- 50
n_chars <- 6
ret <- matrix(rnorm(TT * K) / 100, TT, K)
Z <- array(rnorm(TT * K * n_chars), dim = c(TT, K, n_chars))
fit_ipca <- ipca_est(ret, Z, nfac = 3)
print(fit_ipca)
#> <sdim_fit [ipca]>
#> Observations : 120
#> Characteristics : 6
#> Factors : 3
#> Factor mean : zero
summary(fit_ipca)
#> Instrumented Principal Components Analysis (IPCA)
#> ----------------------------------------
#> Call: ipca_est(ret = ret, Z = Z, nfac = 3)
#>
#> Dimensions
#> ----------------------------------------
#> Observations 120
#> Characteristics 6
#> Factors 3
#> Factor mean zero
#>
#> Eigenvalues
#> ----------------------------------------
#> F1 F2 F3
#> Eigenvalue 0.8952 0.9336 0.8652
#> Var. expl. (%) 33.2300 34.6600 32.1200
# With factor mean specifications
fit_const <- ipca_est(ret, Z, nfac = 3, factor_mean = "constant")
fit_const$mu # time-series mean of each factor
fit_var <- ipca_est(ret, Z, nfac = 3, factor_mean = "VAR1")
fit_var$var_coef # K x K VAR(1) coefficient matrixy <- rnorm(200) # univariate return series
fit_spca <- spca_est(target = y, X = X, nfac = 3)
summary(fit_spca)
#> Scaled PCA (sPCA)
#> ----------------------------------------
#> Call: spca_est(target = y, X = X, nfac = 3)
#>
#> Dimensions
#> ----------------------------------------
#> Observations 200
#> Predictors 20
#> Factors 3
#>
#> Eigenvalues
#> ----------------------------------------
#> F1 F2 F3
#> Eigenvalue 12.3456 8.7654 5.4321
#> Var. expl. (%) 46.73 33.20 20.57
#>
#> OLS slope summary (beta)
#> ----------------------------------------
#> 0% 25% 50% 75% 100%
#> -0.1234 -0.0512 0.0103 0.0634 0.1521If you encounter a bug, please file an issue with a minimal reproducible example on GitHub. For questions, email gabriel.cabreraguzman@postgrad.manchester.ac.uk.
These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.