Objective To investigate the value of back propagation (BP) neural network for recognizing gastric cancer cell. Methods A total of 510 cells was selected from 308 patients. There were 210 gastric adenocarcinoma cells and 300 non-cancer gastric cells. Ten morphological parameters were measured for each cell. These data were randomly divided into two groups: training dataset (A) and test dataset (B). A three-layer BP neural network was built and trained by using dataset A. The network was then tested with dataset A and B.Results For data A, the sensitivity of network was 99%, specificity 99%, positive predictive value 98%, negative predictive value 99%, and accuracy 99%. For data B, the sensitivity of network was 99%, specificity 97%, positive predictive value 96%, negative predictive value 99%, the accuracy 98%. With receiver operator characteristic (ROC) curve evaluation, the area under ROC curve was 0.99.Conclusion The model based on BP neural network is very effective. A BP neural network can be used for effectively recognizing gastric cancer cell.
Most of the existing near-infrared noninvasive blood glucose detection models focus on the relationship between near-infrared absorbance and blood glucose concentration, but do not consider the impact of human physiological state on blood glucose concentration. In order to improve the performance of prediction model, particle swarm optimization (PSO) algorithm was used to train the structure paramters of back propagation (BP) neural network. Moreover, systolic blood pressure, pulse rate, body temperature and 1 550 nm absorbance were introduced as input variables of blood glucose concentration prediction model, and BP neural network was used as prediction model. In order to solve the problem that traditional BP neural network is easy to fall into local optimization, a hybrid model based on PSO-BP was introduced in this paper. The results showed that the prediction effect of PSO-BP model was better than that of traditional BP neural network. The prediction root mean square error and correlation coefficient of ten-fold cross-validation were 0.95 mmol/L and 0.74, respectively. The Clarke error grid analysis results showed that the proportion of model prediction results falling into region A was 84.39%, and the proportion falling into region B was 15.61%, which met the clinical requirements. The model can quickly measure the blood glucose concentration of the subject, and has relatively high accuracy.