Engineering and designing an artificial intelligence tool for pattern classification has promising use in the field of medicine. The data mining approach integrates Decision Trees (DTs); Artificial Neural Networks (ANNs), specifically a Classification-based Multi-Layer Perceptron (MLP); and a MLP ANN with risk stratification. This tool predicted mortality in an adult intensive care unit: sensitivity of 75.0%, specificity of 89.54% and an area under the curve (AUC) of 0.9417 for postoperative cases; non-postoperative cases had a sensitivity of 90.90%, specificity 75.16% and an AUC of 0.8333. Prediction of high-risk preterm birth had a sensitivity of 65.13%, specificity of 84.07% and an AUC of 0.8195 for Parous cases and a sensitivity of 61.08%, specificity of 71.14%, and an AUC of 0.7195 for Nulli-parous cases. The trained integrated tool reduced the complexity while yielding prediction accuracies that exceed current results in the literature. |