com.agenarisk.api.tools

Class SensitivityAnalyser

java.lang.Object

com.agenarisk.api.tools.SensitivityAnalyser

public class SensitivityAnalyser
extends Object

SensitivityAnalyser performs sensitivity analysis of the model for the provided target and sensitivity Nodes.
The input is a JSON configuration. The output includes Tables, Tornado Graphs, Response Curve Graphs.
Discrete and continuous Nodes are supported.
All sensitivity and target Nodes must reside in the same Network.

Author:

Eugene Dementiev

Constructor Summary

Constructors

Constructor and Description
SensitivityAnalyser(Model model, org.json.JSONObject jsonConfig) Constructor for Sensitivity Analysis tool. The Model will be factorised and converted to static taking into account any overriding model settings and observations pre-entered into the selected DataSet. If no DataSet or Network are specified, the first one of each will be used.

Method Summary

Modifier and Type	Method and Description
org.json.JSONArray	buildResponseCurveGraphs() Compiles data for response curve graphs. Only allowed when target node is Numeric Interval. A graph is generated for each sensitivity node and selected summary statistic. Will clip data points outside of lower and upper sensitivity percentile values. X axis is sensitivity node states. Y axis is summary statistic values.
org.json.JSONArray	buildTables() Compiles Sensitivity Analysis data as tables. A table is generated for each sensitivity node. Row per sensitivity state. Column per summary statistic (if the target node is numeric continuous) or target state (if the target node is discrete). First item in a row is a sensitivity state. Positive infinity will appear as "Infinity", negative infinity as "-Infinity" and Double.NaN as "NaN", all wrapped in quotes.
org.json.JSONArray	buildTornadoGraphs() Compiles Sensitivity Analysis data as tornado graphs. Graphs are created for each selected summary statistic.
org.json.JSONObject	getConfig() Compiles an effective configuration of this analysis and returns it as a JSON.
org.json.JSONObject	getFullReport() Generates response curve graphs, tables and tornado graphs. Also includes full JSON configuration for this analysis.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

SensitivityAnalyser

public SensitivityAnalyser(Model model,
                           org.json.JSONObject jsonConfig)
                    throws SensitivityAnalyserException

Constructor for Sensitivity Analysis tool.
The Model will be factorised and converted to static taking into account any overriding model settings and observations pre-entered into the selected DataSet.
If no DataSet or Network are specified, the first one of each will be used.

Parameters:

model - Model to run analysis on

jsonConfig - configuration to override defaults for analysis

Throws:

SensitivityAnalyserException - upon failure

Method Detail

getFullReport

public org.json.JSONObject getFullReport()

Generates response curve graphs, tables and tornado graphs.
Also includes full JSON configuration for this analysis.

Returns:

JSON object of the report

buildTables

public org.json.JSONArray buildTables()

Compiles Sensitivity Analysis data as tables.
A table is generated for each sensitivity node.
Row per sensitivity state.
Column per summary statistic (if the target node is numeric continuous) or target state (if the target node is discrete).
First item in a row is a sensitivity state.
Positive infinity will appear as "Infinity", negative infinity as "-Infinity" and Double.NaN as "NaN", all wrapped in quotes.

Returns:

JSON array of tables

buildTornadoGraphs

public org.json.JSONArray buildTornadoGraphs()

Compiles Sensitivity Analysis data as tornado graphs.
Graphs are created for each selected summary statistic. The graph represents stat(targetNode).
For each sensitivity node there is a bar in the graph.

Returns:

JSON array of tornado graphs

buildResponseCurveGraphs

public org.json.JSONArray buildResponseCurveGraphs()

Compiles data for response curve graphs.
Only allowed when target node is Numeric Interval.
A graph is generated for each sensitivity node and selected summary statistic.
Will clip data points outside of lower and upper sensitivity percentile values.
X axis is sensitivity node states.
Y axis is summary statistic values.

Returns:

JSON array of ROC graphs' data

getConfig

public org.json.JSONObject getConfig()

Compiles an effective configuration of this analysis and returns it as a JSON.

Returns:

JSON object representing effective configuration of this analysis