Bootstrap technique offers the most powerful and efficient implementation tools for the modern multivariate analysis, test of hypotheses, knowledge acquisition, experimental inference.
This paper proposes algorithmic solutions of complex data based problems: Selection of distributions, regression models, discrimination analyses, imputations for missing data, diagnostics, predictions, etc, without the need for using unrealistic, unverifiable assumptions. Advanced bootstrap methods allow a full extraction of knowledge from the data base ad receive responses of interest, i.e. create an automatic generator of inferences. The obtained results define the design of new family of knowledge based intelligent systems.