A comparative Study for Analysis and Classification of the Electronic Passport using Genetic Algorithm and Logistic Regression ( Case study of the Sudanese Electronic Passport)

Abstract

For the purpose of comparison of the traditional statistical methods used in analysing and classifying data with a vast amount of information and modern statistical methods computerized to obtain relationships and models that help in extracting useful information and support for decision making and hence the problem of study in how Employing this field in the organization, analysis and classification of such a large amount of data, and the establishment of a general rule concerning the electronic passport (Sudan model) using the best statistical methods, studied and analyzed the traditional and modern statistical models and methods and the comparison between them and linking The genetic algorithm with some traditional statistical methods was done by using the experimental side of these methods using simulations first to convert them into hypothetical data that simulates real reality in order to arrive at the best way to solve these problems at the lowest cost and shortest possible time, designed A simulation program using the genetic algorithm and through the MATLAB version program (R2017B) and after analyzing the data was compiled using the program of the statistical Package of social Sciences (SPSS 20) The study examined two aspects of the first represented by the theoretical and experimental side, and reviewed the methods Traditional analysis and demonstrate how important it is to apply it in the genetic algorithm to choose the best, this is demonstrated by the second practical aspect that has been designed and applied to the original data of the study through the application of a programmatic model that fits the reality of the data, it was found that the best way is the way of squares Weighted Micro (WLS) results from classical methods and for different sample sizes in estimating the logistic regression on the experimental side, and the greatest possible method (MLE) based on the genetic algorithm emerged as the best method and for all sample sizes, as all methods based on the algorithm exceeded Genetic for classical methods. The study found recommendations using methods based on the genetic algorithm to estimate the parameters of the logistic regression, and the results of the algorithms for data mining can be adopted especially when dealing with a large amount of data, as it can deal with e-governance data Smart and using artificial intelligence algorithms, including the genetic algorithm with cluster analysis algorithms in data mining.