Logistic regression can handle small datasets by using shrinkage methods such as penalized maximum likelihood or Lasso. These techniques reduce regression coefficients, improving model stability and preventing overfitting, which is common in small sample sizes (Steyerberg et al., 2000).
Steyerberg, E., Eijkemans, M., Harrell, F. and Habbema, J. (2000) ‘Prognostic modelling with logistic regression analysis: a comparison of selection and estimation methods in small data sets’, Statistics in medicine, 19(8), pp. 1059-1079.
on theorems big data uses trends to study the models..on small data the best way to try it is using mixed distributions for example Binomial models Bin(n,p) with p dependant of other model p~X(n-1). the reason is on small data the histogram is not static, it changes with any experiment. good day
I have the book Statistical methods in medical research by Armitage and Berry
- Badges
- Science topic
- Similar topics
- Agricultural Science
- Agronomy