Among the most widely used statistical methods in classifyieg data are logistic regression, and discriminatory analyzed, which can be used as
2
linear models in data classification.
There should be a number of assumptions for using analysis discriminatory, the most important among them is that the data of the independent variables are continuous and follow the normal distribution, and the population have common variance matrix. While in logistic regression, there are no conditions concerning the independent variables, as it is encountered to be more accurate than analyzed discriminatory, and it is also simple and flexible, and gives a clear significant explanation and a meaning for describing the relation between both dependent and independent variations.
This study also aims at classifying the data based on three models: the first model is the probability function for the multi-responsive logistic regression. The second model is linear discriminatory function. The third model is the prospected discriminatory function response; The researcher used two models of classification and prediction, which are analyzed discriminatory and logistic regression and made a comparison between them to identify the most important factors that affect the nature of the disease, these factors might be economic, demographic, medical, etc. which are represented in dependent variables, and how it has an effect on the independent variables , besides recognizing which of the two models is better and more accurate in results, and which of them shows the least percentage of fault in classification.