data_density_command#

View page source

Data Density Command: Change the Density for an Integrand#

Syntax#

dismod_at database data_density
dismod_at database data_density integrand_name density_name eta_factor nu

Purpose#

This command is used to change the density used for an integrand during subsequent fits. . If density_id corresponds to binomial, the corresponding meas_std is null and hence the density can’t be changed.

database#

Is an http://www.sqlite.org/sqlite/ database containing the dismod_at input tables which are not modified. If this is the only argument, all the data densities are set back to the values specified by the data_table .

integrand_name#

This is the integrand that we setting the density for.

density_name#

This is the density_name that we are using for the new density, except when the original density is binomial.

eta_factor#

This specifies the value of eta used with this density setting. This value is not used when the density is Linear (You can use the text null in this case). If the density is Log Scaled , the value of eta used with this density is

eta = eta_factor * median ( meas_value )

Here median ( meas_value ) is the median of the meas_value in the data table.

nu#

This specifies the value of nu used with this density setting. This value is only used when the density is students or log_students . (You can use text null is when it is not used.)

data_subset_table#

Only rows of the data_subset_table that correspond to this integrand are modified.

density_id#

The subset table density_id for this integrand is changed to this density .

eta#

If the density is Log Scaled , the subset table eta is set to the value defined above. Otherwise, it is set to null.

nu#

If the density is students or log_students , the subset table nu is set to the value specified above. Otherwise, it is set to null.

Example#

The file user_data_density.py contains an example and test using this command.