Day 28 (April 30)
Announcements
- Calendar
- Lecture on Tuesday 4/30
- Lecture on Tuesday 5/5
- Lecture on Tuesday 5/7
- Next week and the week after is very open if you need to meet.
- Presentation dates/time? My preference for presentations is May 4-8. Please send me an email (thefley@ksu.edu) and provide three 30 min long times/dates that work for you.
- New overview stats paper on physics informed neural networks (link)
- Mechanistic vs. descriptive models
- Physics informed neural networks really blends the two
- Questions/clarifications from journals
- Spatially varying slope models (see here and exam/score example)
- “In the earth quake example you made the choice for IPPP distribution is there a specific methodology or procedure one can use to make selection of distribution.”
- “The earthquake example we have been doing in class, I had always made me wonder why you chose an exponential form instead of simply building the model with distance as the predictor.”
- “What is still not clear to me is the difference between modeling space and time as separate smooth components versus putting latitude, longitude, and time in the same smooth.”
Mechanistic spatio-temporal models