=========================================================================
GFM: Generalized factor model for ultra-high dimensional variableswith mixed types.
GFM is a package for analyzing the (ultra)high dimensional data withmixed-type variables, developed by the Huazhen Lin’s lab. It is not onlycomputationally efficient and scalable to the sample size increment, butalso is capable of choosing the number of factors. In our JASA paper, atwo-step method is proposed to estimate the factor and loading matrix,in which the first step used the alternate maximization (AM) algorithmto obtain initial estimator. In the paper, the information criterion wasprovided to determine the number of factors. Recently, we proposed anoverdispersed generalized factor model (OverGFM) and designed avariational EM algorithm to implement OverGFM. A singular value ratiobased method was provided to determine the number of factors. Inaddition, the estimate from OverGFM can be also used as the initialestimates in the first step for GFMs in our previous JASA paper.
Check out ourJASApaper for alternate maximization and information criterion, and ourPackagevignette for a more complete description of the usage of GFM andOverGFM.
GFM can be used to analyze experimental dataset from different areas,for instance:
Please see our new paper for model details:
To install the the packages ‘GFM’ from ‘Github’, firstly, install the‘remotes’ package.
install.packages("remotes")remotes::install_github("feiyoung/GFM")Or install the the packages “GFM” from ‘CRAN’
install.packages("GFM")For an example of typical GFM usage, please see ourPackagevignette for a demonstration and overview of the functions includedin GFM.
GFM version 1.2.1 (2023-08-10)
The functionoverdispersedGFM() that implements theoverdispersed generalized factor model is added. In addition, thefunctionOverGFMchooseFacNumber() is added, whichimplements singular value ratio (SVR) based method to select the numberof factors.