Version History

What is HB-Reg?

HB-Reg is a system for hierarchical Bayes estimation of general regression-based problems.
You may e-mail support@sawtoothsoftware.com if you have questions relating to information on this web page.
We strongly suggest using the latest version of HB-Reg.

Expand All

Version 4 History

Changes

  1. Beginning in this version, the jump size does not change after burn-in iterations are completed. This change is due to feedback from academics and HB experts.
  2. Updated terms of service.
Read More

Fixes

  1. Includes various minor fixes.
Read More

Fixes

  1. Fixed display issue with covariates when using high DPI settings.
  2. Fixed problem copying models to clipboard.
Read More

Changes

  1. This version allows users to get licenses by signing into Sawtooth Account.
Read More

Fixes

  1. This version fixes an issue with respondent filtering. This can occur if the locale set in the software used to read CSV files is different than the machine’s locale. Decimal values are not converted to numbers correctly.
Read More
  1. This version has a new feature which allows for the import of model specifications from a CSV file.
  2. If a variable was coded linear and did not have initialized category values, they could receive an error in estimation that no respondents were included.
Read More
  1. Modified the file reader to better identify missing values.
Read More
  1. When saving draws, the point estimates file would show utilities getting larger for each respondent. This was because HB-Reg was not resetting the sum of the draws to zero between each respondent.
Read More
  1. Utility archive files would become corrupt if the draws file exceeded 2GB in size. The software will now allow draw files up to 8GB.
  2. The estimation time contained in the HB estimation log was incorrect.
Read More
  1. Bug: When using multiple dependent variables, the grid for selecting covariates could become disabled and would not re-enable.
Read More
  1. Bug: Using constraints on a model that involved independent variables coded as 'excluded' could result in an 'index out of range' error.
Read More
  1. Bug: v4.0.0 did not correctly recognize Academic Research licenses.
  2. Bug: Users could see an error when creating a new model for a dependent variable.
  3. Bug: Viewing a demographic file would cause a crash if the file was open in another editor such as Excel.
  4. Bug: Models would attempt to estimate even if all independent variables were set to Excluded.
Read More
  1. Version 4 features a new user interface, borrowed extensively from the Sawtooth Software's MBC (Menu-Based Choice) software product.
  2. HB-Reg 4 can automatically code Categorical Independent variables as dummy-coded, linear, or log-linear. This removes extra data processing requirements, making it easy for you to switch back and forth between different independent variable coding schemes.
  3. HB-Reg 4 automatically builds a "proper prior covariance matrix" when you specify a categorical independent variable (when dummy-coding is activated). This is the same procedure as recommended by leading Bayesian academic Peter Lenk that is also used in our CBC/HB software.
  4. Covariates may now be included in HB runs. For more information on covariates in HB, please see the white paper entitled: Application of Covariates within Sawtooth Software's CBC/HB Program: Theory and Practical Example (2009). This paper may be downloaded from our Technical Papers library on our website: www.sawtoothsoftware.com.
  5. Ability to run multiple HB runs simultaneously, utilizing multiple cores of your machine's processor.
  6. Big gains in speed for exceptionally large datasets (>128MB) due to an improved memory caching technique. We have an enormous HB-Reg dataset for testing (250MB) that experienced a 20x increase in speed under HB-Reg v4 compared to v3.
Read More

Version 3 History

  1. Feature: Without a license HB-Reg now operates in demonstration mode. The separate demo download has been removed.
  2. Change: A list of the independent variables used is saved to the log file.
  3. Bug: Extremely large files caused problems sorting draws.
Read More

19 April 2007

  1. Change: Updated to work with the .NET Framework 2.0 and Windows Vista.
  2. Bug: Numbers typed in the 'Independent variable number(s)' field and later removed would reappear the next time the study was loaded.
Read More

08 September 2004

  1. Bug: If more than 1000 saved draws were requested, the .LOG and .PNT would be truncated. This did not affect the quality of the results.
Read More

02 January 2004

  1. Change: The range for the prior variance has been changed to (0.001 - 100,000,000.0).
Read More

15 December 2003

Initial Release

Read More