An implementation of the feature Selection procedure by Partitioning the entire Solution Paths (namely SPSP) to identify the relevant features rather than using a single tuning parameter. By utilizing the entire solution paths, this procedure can obtain better selection accuracy than the commonly used approach of selecting only one tuning parameter based on existing criteria, cross-validation (CV), generalized CV, AIC, BIC, and EBIC (Liu, Y., & Wang, P. (2018) https://doi.org/10.1214/18-EJS1434). It is more stable and accurate (low false positive and false negative rates) than other variable selection approaches. In addition, it can be flexibly coupled with the solution paths of Lasso, adaptive Lasso, SCAD, MCP, ridge regression, and other penalized estimators.

The `SPSP`

package is currently available on SPSP CRAN.

`SPSP`

development version from GitHub (recommended)```
# Install the development version from GitHub
if (!requireNamespace("devtools")) install.packages("devtools")
::install_github("XiaoruiZhu/SPSP") devtools
```

`SPSP`

from the
CRAN```
# Install from CRAN
install.packages("SPSP")
```

The user-friendly function SPSP() conducts the selection by
Partitioning the Solution Paths (the SPSP procedure) to selects the
relevant predictors. The user only needs to specify the independent
variables matrix, response, family, and a penalized method that can
generate the solution paths, for example, Lasso, adaptive Lasso, SCAD,
MCP, ridge regression. The embedded selection methods in this package
can be called using `fitfun.SP = lasso.glmnet`

. Currently,
six methods are included: `lasso.glmnet`

,
`adalasso.glmnet`

, `adalassoCV.glmnet`

,
`SCAD.ncvreg`

, `MCP.ncvreg`

, and
`ridge.glmnet`

.

The following example shows the R codes:

```
library(SPSP)
data(HihgDim)
library(glmnet)
<- as.matrix(HighDim[,-1])
x <- HighDim[,1]
y
# SPSP + lasso
<- SPSP(x = x, y = y, family = "gaussian", fitfun.SP = lasso.glmnet,
spsp_lasso_1 init = 1, standardize = FALSE, intercept = FALSE)
head(spsp_lasso_1$nonzero)
head(spsp_lasso_1$beta_SPSP)
# SPSP + adalasso
<- SPSP(x = x, y = y, family = "gaussian", fitfun.SP = adalasso.glmnet,
spsp_adalasso_5 init = 5, standardize = T, intercept = FALSE)
head(spsp_adalasso_5$nonzero)
head(spsp_adalasso_5$beta_SPSP)
# SPSP + SCAD
<- SPSP(x = x, y = y, family = "gaussian", fitfun.SP = SCAD.ncvreg,
spsp_scad_5 init = 5, standardize = T, intercept = FALSE)
head(spsp_scad_5$nonzero)
head(spsp_scad_5$beta_SPSP)
```

Liu, Y., & Wang, P. (2018). Selection by partitioning the
solution paths. *Electronic Journal of Statistics*, 12(1),
1988-2017. <10.1214/18-EJS1434>