Evaluation of human liver condition Self-organizing map and feed forward backpropagation technique

Serum Glutamic Pyruvate Transferase (SGPT) is an enzyme that used as a medical standard to evaluate health of human liver. The only method for measuring the amount of SGPT is through blood sampling. This paper introduces a new approach to measure the level on SGPT using body composition and home-used measuring tool. The self-organizing map was use as clustering tool and feature extraction tool. The prediction model was synthesized by using multi-layered feedforward neural network. The accuracy of the presented approach is reaching an impressive rate of 91%–97%, depending on the network structure and training features.