Here
run() method gets a 3-d array of data and a
2-d array of coefficients as arguments and returns a 2-d array of utilities.
A child class of linear_utilities, called
linear_utilities_diagnose,
serves analyzing changes in utilities
due to changes in the data. For each variable set to its 5%- and 95%-quantile, respectively, while keeping
the remaining variables at their median, it runs the linear_utilities module. Results
are stored in a file. First line of the file corresponds to the 5%-quantile, second line to the 95%-quantile,
and third line contains the difference between lines one and two. The name of the file
can be passed to the run() method in its argument resources as a dictionary key
'utilities_diagnose_file'. Default name is 'util'.