Week 1 - Linear Selection Function

Linear selection function

W(z)=α+βz

z → numerical value of a trait

α → W-intercept

β → slope

• Describes the relative fitness, W, of an organism with phenotypic trait z.

• W is dependent variable, z is independent variable

• W values closer to 0 → low fitness

• W values closer to 1 → high fitness

The maximum range of W is [0,1]

Interpolation

predicting a value within the domain and range of data

Extrapolation

Predicting a value outside of the domain and range of the data

Linear regression

(AKA “line of best fit”) A linear function that most closely represents the data when there is an observable linear pattern

Correlation coefficient

Describes the strength and relationship of data

represented by “r”

r>0 → positive/increasing relationship

r<0 → negative/decreasing relationship

• the closer r is to 0, the more scattered the data → meaning that the line of best fit/linear regression would not have good predictive power

• r closer to -1 or 1 → data is more tightly packed together, less scattered → meaning the line of best fit/linear regression has good predictive power