Future direction of Ngene
Posted: Mon Feb 18, 2013 4:53 pmDear Fellow Ngener’s
We understand that many of you are keen to see what plans we have for future versions of Ngene in terms of features. Whilst it is a little early to be definitive, below is a list of current topics we are working on. We hope to bring these into version 2, however we are at least 12 months off this (at a minimum)
1. Optimal Best Worst designs (Cases 1, 2, and 3).
2. Availability and variable choice set designs.
3. Allowing priors for the ignoring of attributes (an extension to modelling average – think information processing strategies).
4. Lots of new types of draws.
5. More design algorithms including genetic algorithms.
6. More flexibility to control things such as attribute level balance.
7. More types of non-linear coding structures beyond dummy and effects coding.
8. Better ways of handling dominance.
9. Speed improvements for MMNL type designs.
10. New model types, including NL, LCM and a few more.
11. Designs allowing for non-linear utility functions.
We have already made significant progress on 1-7, and have developed all the theories for 8-11. That said, we still have a huge amount of testing to do on all items, and significant programming challenges to overcome.
John
We understand that many of you are keen to see what plans we have for future versions of Ngene in terms of features. Whilst it is a little early to be definitive, below is a list of current topics we are working on. We hope to bring these into version 2, however we are at least 12 months off this (at a minimum)
1. Optimal Best Worst designs (Cases 1, 2, and 3).
2. Availability and variable choice set designs.
3. Allowing priors for the ignoring of attributes (an extension to modelling average – think information processing strategies).
4. Lots of new types of draws.
5. More design algorithms including genetic algorithms.
6. More flexibility to control things such as attribute level balance.
7. More types of non-linear coding structures beyond dummy and effects coding.
8. Better ways of handling dominance.
9. Speed improvements for MMNL type designs.
10. New model types, including NL, LCM and a few more.
11. Designs allowing for non-linear utility functions.
We have already made significant progress on 1-7, and have developed all the theories for 8-11. That said, we still have a huge amount of testing to do on all items, and significant programming challenges to overcome.
John
