Mining Data and Building Predictive Models - Virtual JMP Day Costa Rica - April 1, 2020 - McCormackTransforming Data to Make Better Predictions webinar mentioned in this presentation.
<script>LITHIUM.OoyalaPlayer.addVideo('https:\/\/player.ooyala.com\/static\/v4\/production\/', 'lia-vid-t1dHFjajE6yZdfj6JIKEmqprFe-Sx_47w1600h900r200', 't1dHFjajE6yZdfj6JIKEmqprFe-Sx_47', {"pcode":"ltbGkyOjMl3V3pZ3Trn6T2z1_rjF","playerBrandingId":"ODI0MmQ3NjNhYWVjODliZTgzY2ZkMDdi","width":"1600px","height":"900px"});(view in My Videos)
What would be resources #1 to learn more about this topic in particular?
A: Great question lots of options. Here is a place that has a bunch of videos. https://www.jmp.com/en_us/events/mastering/application-areas/statistics-predictive-modeling-and-data-mining.html
Re: Mining Data and Building Predictive Models - Virtual JMP Day Costa Rica - April 1, 2020 - McCormackQ&A from the online presentation:
Q: Are there good normalization options available in JMP - or is the user expected to do this through a formula?
A: check out this video on the topic:
https://www.jmp.com/en_us/events/ondemand/mastering-jmp/transforming-data.html
A: If you are interested in Generalized Regression, this is a good talk too. In GenReg, you need to specify a distribution of the Y. This is a great talk on that.
https://community.jmp.com/t5/Discovery-Summit-Tucson-2019/Not-Quite-Normal-Choosing-the-Best-Distribution-for-Modeling/ta-p/223356
Re: Mining Data and Building Predictive Models - Virtual JMP Day Costa Rica - April 1, 2020 - McCormackQ&A from the online presentation:
Q: Where to find out more about the K-nearest neighbor modeling options?
A: Modeling K-nearest Neighbor analysis is here: https://www.jmp.com/support/help/14/example-of-k-nearest-neighbors-with-categorical.shtml
A: If you want to look at k-nearest neighbors in multivariate analysis:
https://www.jmp.com/support/help/14-2/multivariate-k-nearest-neighbor-outliers.shtml
Re: Mining Data and Building Predictive Models - Virtual JMP Day Costa Rica - April 1, 2020 - McCormackQ&A from the online presentation:
Q: How do I fit multi-dimentional data? For example, I have an image, x,y is pixel coordinate. I want to create a model using x,y and parameter color to create a model, so that I can predict values at different x,y location.
A: You could have x and y into the Y (response) box and build a model for both outcomes. The profilers would stack and you can optimize both x and y models at the same time. Separately, you might want to use a mixed model that includes spatial dependency.
Q: Where can I learn about mixed models?
A: https://community.jmp.com/t5/JMP-Blog/JMP-Pro-for-linear-mixed-models-Part-1/ba-p/30433
A: https://www.jmp.com/en_us/events/ondemand/mastering-jmp/linear-mixed-models-using-jmp-pro.html
A: also https://www.jmp.com/support/help/14/mixed-models.shtml#
Re: Mining Data and Building Predictive Models - Virtual JMP Day Costa Rica - April 1, 2020 - McCormackQ&A from the online presentation:
Q: Is the validation data splits the same for each model, or is it changing from model to model?
A: In this case he is keeping them the same
A: You can change them to see if it improve validation fit. Keeping it the same allows for a true comparison at the end
