augsynth: Fit Augmented SCM

View source: R/augsynth_pre.R

augsynthR Documentation

Fit Augmented SCM

Description

Fit Augmented SCM

Usage

augsynth(form, unit, time, data, t_int = NULL, ...)

Arguments

form

outcome ~ treatment | auxillary covariates

unit

Name of unit column

time

Name of time column

data

Panel data as dataframe

t_int

Time of intervention (used for single-period treatment only)

...

Optional arguments

  • Single period augsynth with/without multiple outcomes

    • "progfunc"What function to use to impute control outcomes: Ridge=Ridge regression (allows for standard errors), None=No outcome model, EN=Elastic Net, RF=Random Forest, GSYN=gSynth, MCP=MCPanel, CITS=CITS, CausalImpact=Bayesian structural time series with CausalImpact, seq2seq=Sequence to sequence learning with feedforward nets

    • "scm"Whether the SCM weighting function is used

    • "fixedeff"Whether to include a unit fixed effect, default F

    • "cov_agg"Covariate aggregation functions, if NULL then use mean with NAs omitted

  • Multi period (staggered) augsynth

    • "relative"Whether to compute balance by relative time

    • "n_leads"How long past treatment effects should be estimated for

    • "n_lags"Number of pre-treatment periods to balance, default is to balance all periods

    • "alpha"Fraction of balance for individual balance

    • "lambda"Regularization hyperparameter, default = 0

    • "force"Include "none", "unit", "time", "two-way" fixed effects. Default: "two-way"

    • "n_factors"Number of factors for interactive fixed effects, default does CV

Value

augsynth object that contains:

  • "weights"weights

  • "data"Panel data as matrices


ebenmichael/augsynth documentation built on Sept. 9, 2024, 3:29 p.m.