Storing output from quantile regressions

Goals

In our previous quantileFit tutorials, we have used the call command to print results directly to the screen. This is convenient for quickly viewing results but you may wish to store your results for later use.

This tutorial introduces the use of the qFitOut structure to store results from quantile regressions. In this tutorial we will learn:

  • How to declare an instance of the qFitOut structure for use.
  • What members are contained in the qFitOut structure.
  • How to view members in the qFitOut structure

Declaring the structure

Prior to use, a structure must be declared using the struct keyword :

struct qFitOut qOut;

Once declared, the structure will be visible in the symbol editor:

quantile regression output structure

Members in the qFitOut structure

The qFitOut structure holds five members:


qOut.beta
Matrix, the estimated regression coefficients. Estimates for separate quantiles are contained in separate columns.
qOut.u_plus
Matrix, the positive part of residuals. Residuals for separate quantiles are contained in separate columns.
qOut.u_minus
Matrix, the negative part of residuals. Residuals for separate quantiles are contained in separate columns.
qOut.ci
Array, bootstrapped confidence intervals. Confidence intervals for separate quantiles are stored in separate planes. These are only computed if bootstrap computations is specified.
qOut.se
Vector, bootstrapped standard errors. Standard errors for separate quantiles are stored in separate columns. These are only computed if bootstrap computations is specified.

These members are filled by the quantileFit procedure:

// Declare structure
struct qFitOut qOut;

// Estimate model
qOut = quantileFit(y, x);

Viewing members of the output structure

The members of the qFitOut structure can be viewed three different ways:

  1. Navigating to the data page and double-clicking on a structure member in the Active Symbol Tree. viewing a structure in the data editor
  2. Placing the cursor anywhere on the structure name in a program file and using Ctrl+E to open a floating Symbol Editor. quantile regression floating symbol editor
  3. Print the member contents to the output screen using the print command.
    print qOut.beta;
      -11.676810        4.7151959        11.997397        21.834192
       1.6789851        2.1301881        2.1965671        2.5095988 

Conclusion

Congratulations! After completing this tutorial you should understand how to store and view the output from the quantileFit procedure.

The next tutorial examines how to change variable names in the quantileFit output table using the qFit.varnames member.

Have a Specific Question?

Get a real answer from a real person

Need Support?

Get help from our friendly experts.

Try GAUSS for 14 days for FREE

See what GAUSS can do for your data

© Aptech Systems, Inc. All rights reserved.

Privacy Policy