Formula Interface for Spending Models

Fit latent Gamma-Gamma model for customer spending with a formula interface

spending(formula, data, optimx.args = list(), verbose = TRUE)

Arguments

formula

Formula specifying the model to be fit. See Details.

data

Either a clv.data object or a data.frame containing transaction data on which the model specified in formula will be fit.

optimx.args

Additional arguments to control the optimization which are forwarded to optimx::optimx. If multiple optimization methods are specified, only the result of the last method is further processed.

verbose

Show details about the running of the function.

Value

Returns an object of the respective model which was fit.

See also

Spending models for inputs: gg.

latentAttrition to fit latent attrition models with a formula interface

Examples

# \donttest{

data("cdnow")
clv.cdnow <- clvdata(data.transactions = cdnow, date.format="ymd",
                     time.unit = "weeks")

# Fit gg
spending(~gg(), data=clv.cdnow)
#> Starting estimation...
#> Estimation finished!
#> Gamma-Gamma Model
#> 
#> Call:
#> spending(formula = ~gg(), data = clv.cdnow)
#> 
#> Coefficients:
#>      p       q   gamma  
#>  4.685   4.188  24.369  
#> KKT1: TRUE 
#> KKT2: TRUE 

# Fit gg with start params
spending(~gg(start.params.model=c(p=0.5, q=15, gamma=2)),
         data=clv.cdnow)
#> Starting estimation...
#> Estimation finished!
#> Gamma-Gamma Model
#> 
#> Call:
#> spending(formula = ~gg(start.params.model = c(p = 0.5, q = 15, 
#>     gamma = 2)), data = clv.cdnow)
#> 
#> Coefficients:
#>      p       q   gamma  
#>  4.683   4.190  24.391  
#> KKT1: TRUE 
#> KKT2: TRUE 

# Fit gg, do not remove first transaction
spending(~gg(remove.first.transaction=FALSE), data=clv.cdnow)
#> Starting estimation...
#> Estimation finished!
#> Gamma-Gamma Model
#> 
#> Call:
#> spending(formula = ~gg(remove.first.transaction = FALSE), data = clv.cdnow)
#> 
#> Coefficients:
#>      p       q   gamma  
#>  7.060   3.752  12.894  
#> KKT1: TRUE 
#> KKT2: TRUE 
# same, abreviate parameters
spending(~gg(remo=F), data=clv.cdnow)
#> Starting estimation...
#> Estimation finished!
#> Gamma-Gamma Model
#> 
#> Call:
#> spending(formula = ~gg(remo = F), data = clv.cdnow)
#> 
#> Coefficients:
#>      p       q   gamma  
#>  7.060   3.752  12.894  
#> KKT1: TRUE 
#> KKT2: TRUE 

# Fit gg on given data.frame transaction data, no split
spending(data()~gg(), data=cdnow)
#> Starting estimation...
#> Estimation finished!
#> Gamma-Gamma Model
#> 
#> Call:
#> spending(formula = data() ~ gg(), data = cdnow)
#> 
#> Coefficients:
#>      p       q   gamma  
#>  4.685   4.188  24.369  
#> KKT1: TRUE 
#> KKT2: TRUE 

# Fit gg on given data.frame, split after 39 periods
spending(data(split=39)~gg(), data=cdnow)
#> Starting estimation...
#> Estimation finished!
#> Gamma-Gamma Model
#> 
#> Call:
#> spending(formula = data(split = 39) ~ gg(), data = cdnow)
#> 
#> Coefficients:
#>      p       q   gamma  
#>  6.238   3.746  15.478  
#> KKT1: TRUE 
#> KKT2: TRUE 
# same but also give date format and period definition
spending(data(split=39, format=ymd, unit=w)~gg(), data=cdnow)
#> Starting estimation...
#> Estimation finished!
#> Gamma-Gamma Model
#> 
#> Call:
#> spending(formula = data(split = 39, format = ymd, unit = w) ~ 
#>     gg(), data = cdnow)
#> 
#> Coefficients:
#>      p       q   gamma  
#>  6.238   3.746  15.478  
#> KKT1: TRUE 
#> KKT2: TRUE 

## No covariate may be selected or covariate data.frame may be
## given because currently no spending model uses covariates

# }