The objective of this study is to identify the determinants of policy bank lending in the context of
Yemen. To achieve this goal, attempt was made to model the probability of lending investment
companies Yemenis through a number of explanatory variables included in the multivariate model
(model discriminant analysis) and a model of artificial intelligence (artificial Neural Network). To
empirically study this problem, forty (40) cases of actual investment credit were collected from eight
banks in Yemenis (28 and 12 cases of agreement of rejection). The two models have highlighted the
importance of non-financial factors including the guarantees (in the form of real estate) with the
registration of these guarantees to finance revenue or conservation land by type of guarantee. Also, it
was noted that the analysis by artificial neural networks allowed us to have a clearer and precise
element that can predict the credit decision. Comparison of the two models in terms of predictability in
lending shows the superiority of the performance art "neuronal" compared to discriminant analysis.
Keywords: Financial statements, bank credit, financial and non-financial ration, Yemen.