ssb-model-solver 1.4.1

Class to define, block analyse and solve dynamic and algebraic models numerically

SSB Model Solver

ModelSolver is a class that defines, block analyses and solves dynamic and algebraic models numerically. See documentation for detailed information about theory and implementation of the class.

Opprettet av: Magnus Kvåle Helliesen [e-mail]

Features

ModelSolver is a Python class. It defines, analyses and solves dynamic algebraic model with lots of equations.

The package is imported using

import model_solver as ms

Usage is

model = ms.ModelSolver(equations, endogenous)

where equations are equations and endogenous are endogenous variables, both stored as strings in lists.

Built with

ModelSolver uses the following packages

• NumPy
• NetworkX
• Pandas
• SymEngine
• Numba
• collections
• functools
• Matplotlib

Example of use

Let equations = ['x+y = 1', 'x-y = 2'] and endogenous = ['x', 'y'], then the model class is initialized by

model = ms.ModelSolver(equations, endogenous)

When initialized, the class reads in the equations, analyzes them for any lags, before it block analyzes it to find the smalles model blocks that must be solved simultaneously. Note that ModelSolver is not case sensitive, such that 'x' and 'X' are the same, both in equations, lists and dataframe (below).

When the class is finished initializing, the user can call the following methods:

• solution = model.solve(dataframe) where dataframe is a Pandas dataframe containing initial values for the endogenous variables and values for the exogenous variables. solution is a dataframe with same dimensions as dataframe containing the solutions for the endogenous variables.
• model.switch_endo_vars(old_endo_var, new_endo_var) switches the endogenous variables old_endo_var for new_endo_var.
• model.describe() writes out information about the model: the number of blocks, the size of the blocks etc.
• model.find_endo_var('var') returns the block number in which var is solved for.
• model.show_block(block_number) returns information about the block: endogenous variables, predetermined variables and equations.
• model.show_blocks() returns information about all blocks.
• model.trace_to_exog_vars(block_nunber) traces back to the exogenous variables that may affect the block.
• model.trace_to_exog_vals(block_nunber, period_index) traces back to the exogenous variable values for the period.
• model.draw_blockwise_graph(variable, maximum_ancestor_generations, maximum_decendants_generations) where variable is a variable of interest, and maximum_ancestor_generations and maximum_decendants_generationsare non-negative integers that governs the number of generations before and after the variable to be graphed. The output is a HTML-file with a relational graph.
• model.sensitivity(block_nunber, period_index[, method='std', exog_subset=None]) analyses the sensitivity of the endogenous variable in the block with respect to the exogenous variabels that determine the solution for the period.

Installation

You can install SSB Model Solver via pip from PyPI:

pip install ssb-model-solver

Usage

Please see the Reference Guide for details.

Contributing

Contributions are very welcome. To learn more, see the Contributor Guide.

License

Distributed under the terms of the MIT license, SSB Model Solver is free and open source software.

Issues

If you encounter any problems, please file an issue along with a detailed description.

Credits

This project was generated from Statistics Norway's SSB PyPI Template.