摘要 |
The present invention relates to a computer implemented process for developi ng a model which predicts the value of a single dependent variable based on the value of at least one independent variable (Figure 6, 100, 102, 104, 106, 10 8, 110, 112). The process comprises the steps of creating a dataset containing a plurality of observations each containing a value for the dependent variable and values for the at least one independent variable, creating from the dataset a plurality of original chromosomes each comprising a possible predictive model, developing a quantitative fitness measure for each chromosome, and creating a new generation of chromosomes by selecting a numb er of the original chromosomes based upon the fitness measures, crossing the selected original chromosomes by at least one of a cloning and a pure (standard) crossover technique, and mutating the crossed chromosomes. A syst em for carrying out the process of the present invention is also described.
|