_apop_model Struct Reference

Data Fields

• char name [101]
• int vbase
• int m1base
• int m2base
• apop_settings_type * settings
• apop_data * parameters
• apop_data * expected
• apop_data * covariance
• double llikelihood
• int status
• int prepared
• apop_data * data
• apop_model *(* estimate )(apop_data *data, apop_model *params)
• double(* p )(apop_data *d, apop_model *params)
• double(* log_likelihood )(apop_data *d, apop_model *params)
• void(* score )(apop_data *d, gsl_vector *gradient, apop_model *params)
• double(* constraint )(apop_data *data, apop_model *params)
• apop_data *(* expected_value )(apop_data *d, apop_model *params)
• void(* draw )(double *out, gsl_rng *r, apop_model *params)
• void(* prep )(apop_data *data, apop_model *params)
• void(* print )(apop_model *params)
• void * more
• size_t more_size

---

Detailed Description

The elements of the apop_model type.

---

Field Documentation

apop_data* _apop_model::covariance

The variance-covariance matrix.

apop_data* _apop_model::expected

An apop_data structure with three columns. If this is a model with a single dependent and lots of independent vars, then the first column is the actual data. Let our model be . Then the second column is the predicted values: , and the third column is the residuals: . The third column is therefore always the first minus the second, and this is probably how that column was calculated internally. There is thus currently no way to get just the predicted but not the residuals or vice versa.

int _apop_model::m2base

The size of the parameter set. If a dimension is -1, then use yourdata->matrix->size2. For anything more complex, allocate the parameter set in the prep method.

apop_data* _apop_model::parameters

The vector of coefficients or parameters estimated by the model.

int _apop_model::status

The return status from the estimate that had populated this apop_model, if any

---

The documentation for this struct was generated from the following file:

• types.h