Evaluation of an artificial neural network to ascertain why there is a high incidence of hepatitis B in the Chinese population after vaccination

Establishment of an analysis model based on measurement of hepatitis B viral infection serum markers

BACKGROUND

Using serum markers of hepatitis B virus (HBV) infection, we aimed to develop a quantitative model that explains the complicated immune response to this infection.

METHODS

Serum samples from HBV-infected patients were randomly selected and divided into groups based on HBV-DNA positivity or negativity. Quantitative markers of HBV were measured. Formulae for Antibody index (IAb) [(anti-HBs * 1/anti-HBe + anti-HBs * 1/anti-HBc + 1/anti-HBc * 1/anti-HBe)^0.5] and Antigen index (IAg) [(HBsAg * HBeAg)^0.5] were introduced.

RESULTS

IAg values were statistically higher (p < 0.05) in the HBV-DNA-positive group than in the -negative group, but no statistically significant difference in IAb values was observed. When IAb values were > 50, IAg values were mostly < 250; when IAg values were > 250, IAb values were mostly < 50.

CONCLUSION

IAb and IAg values can efficiently reflect the status of immune response to HBV and may be suitable for assessment of the infection process and the possible outcome of infection.

Rapid discrimination between red and white ginseng based on unique mass-spectrometric features

Red ginseng (RG) and white ginseng (WG), two processed products of Panax ginseng C. A. Meyer, are in high demand due to their unique features. In this study, some of these unique features were identified and confirmed as biomarkers of RG by using ultra-high-performance liquid chromatography-mass spectrometry, data mining, support vector machine, and artificial neural network. Principal component analysis showed clear separation between the RG and WG extracts, indicating the presence of potential discriminators. In addition, 20 features that are dominant in RG were found by data mining. Samples of Panax quinquefolium (PQ) and Panax notoginseng (PN), close relatives of Panax ginseng C.A.Meyer, were investigated and it was found that 17 features which were absent in PQ and PN samples, were present in RG and WG. Five of these markers were identified as nitrogen-containing compounds that have not been previously reported. Finally, we found that RG can be identified among different ginseng medicinal herbs including RG, WG, PQ, and PN samples, by loading four feature markers corresponding to nitrogen-containing compounds into a discriminating model, based on a support vector machine or an artificial neural network. Thus, this study provides an efficient tool to identify RG during pharmacological research.

Application of knowledge discovery process on the prediction of stroke

OBJECTIVE

Stroke is a prominent life-threatening disease in the world. The current study was performed to predict the outcome of stroke using knowledge discovery process (KDP) methods, artificial neural networks (ANN) and support vector machine (SVM) models.

MATERIALS AND METHODS

The records of 297 (130 sick and 167 healthy) individuals were acquired from the databases of the department of emergency medicine. Nine predictors (coronary artery disease, diabetes mellitus, hypertension, history of cerebrovascular disease, atrial fibrillation, smoking, the findings of carotid Doppler ultrasonography [normal, plaque, plaque+stenosis≥50%], the levels of cholesterol and C-reactive protein) were used for predicting the stroke. Feature selection based on the Cramer's V test was carried out for reducing the predictors. Multilayer perceptron (MLP) ANN and SVM with radial basis function (RBF) kernel were used for the prediction based on the selected predictors.

RESULTS

The accuracy values were 81.82% for ANN and 80.38% for SVM in the training dataset (n=209), and 85.9% for ANN and 84.62% for SVM in the testing dataset (n=78), respectively. ANN and SVM models yielded area under curve (AUC) values of 0.905 and 0.899 in the training dataset, and 0.928 and 0.91 in the testing dataset, consecutively.

CONCLUSION

The findings of the current study pointed out that ANN had more predictive performance when compared with SVM in predicting stroke. The proposed ANN model would be useful when making clinical decisions regarding stroke.