Towards Digital Sustainability: Integrating Canonical Correlation with Artificial Neural Network

The study aims to use an artificial neural network to analyze the canonical correlation by determining the optimal weights, as well as comparing the results of the correct correlation and the canonical correlation using the artificial neural network. And comparing the results of the correlation between two sets of variables using the canonical correlation and the legal correlation using the artificial neural network to show the extent to which the results of the artificial neural networks match the results of the statistical model to adopt the artificial neural network model as another means of conducting the canonical correlation analysis. The study concluded that the group represented by (the age of the husband, the age of the wife, the educational level of the husband, the educational level of the wife, the husband’s occupation, the wife’s occupation, the standard of living of the family, the presence of kinship between the spouses, independent housing, the husband’s smoking, the husband’s consumption of alcoholic beverages) had a clear and significant effect on the variables of the second group, represented by (the variable of the duration of marriage, the variable of the number of children). The first function also proved its high significance according to the value of the canonical correlation coefficient, which helps in making the decision that the first function is the function that can be relied upon in interpreting the relationship between the two components, meaning that the function the first is reliable in analyzing the relationship between the first and second group of variables, as well as using it in prediction. The study also concluded that the best neural network model is the third model, as it has the lowest value for the criteria for comparing error measures and the highest coefficient of determination, meaning that the explanatory power of this model reached 97% of what can be relied upon in building canonical variables and prediction.