How do I use JMP Pro to do machine learning predictions on a dataset with predictors that include latitudes/longitudes (i.e. a geospatial dataset)?Hi - I would like to do a machine learning algorithm in JMP Pro 17 to predict Y based on 7 variables (X1, X2, ...X7), two of which are spatial coordinates (latitude and longitude) and one of which is a date. I would use a portion of my records as a training set, and a portion as my verification/testing set. Is JMP Pro equipped to machine learn on this geospatial (lat/long) dataset? And, if so, could you recommend a good starting point or test type for my workflow? I am using JMP Pro 18.
Thanks in advance!
Predictive Modeling and Machine LearningRe: How do I use JMP Pro to do machine learning predictions on a dataset with predictors that include latitudes/longitudes (i.e. a geospatial dataset)?Geospatial modeling or kriging often uses Gaussian Process models. They are available under Analyze > Specialized Modeling.
Re: How do I use JMP Pro to do machine learning predictions on a dataset with predictors that include latitudes/longitudes (i.e. a geospatial dataset)?Thank you, Mark. I experimented with this yesterday but realized that I need to better understand this type of model (Gaussian Process) to properly interpret the results. I'm curious though if this is the only geospatial modeling option in JMP Pro? I have found very little in the JMP literature on the use of geospatial coordinates for predictions, and yet this seems like such a basic component to understanding, for example, environmental data. I'm probably missing something right in front of me!
Re: How do I use JMP Pro to do machine learning predictions on a dataset with predictors that include latitudes/longitudes (i.e. a geospatial dataset)?This is a broad question within a broad field. JMP does not strive to excel in this field, but many modeling platforms and add-ins could be applied. It makes sense to me to start with researching the field and then check to see if JMP provides appropriate methods.
