Numerical Modeling Techniques GEOG 475 TAMU

  • Empirical curve fitting

    • Here is a curve that fits the trend, so lets just use the equation on the curve

    • Basically multiple regression

  • Linear regression

    • Variance in independent variable to predict the variance in the dependent variable

    • Can predict y by doing x

    • Can use 50-60 x-s to predict y

    • Has to be trained

    • Parametric stat two variables

    • Stronger the prob, stronger relationship, no relationship is no prob

  • Multiple regression

    • Parametric

    • Lots of variables 

    • Has to be linear

    • Can transform to polynomial 

  • Geographically-weighted regression

    • Prediction at local scale in multiple places to come up with what the __________ should be based on many locations

    • parameteric

  • Logistic regression

    • Exists or doesn't exist

    • Only two options

    • X or y; yes or no

    • 1s and 0s and prob of 1s and 0s

    • Based on linear regression

      • Has to be trained

  • Cellular Automata

    • Basically AI

    • Rule-based 

  • Suitability-site analysis (Criteria-based modeling)

    • Where is best for x (ranked)

    • How suitable is x for these uses/criteria

    • Empirical 

  • Agent-based modeling

    • Model individuals or animals, how they interact, how to make decisions, and how the population responds

    • Numerical modeling