Probabilistic neural network for aroyl compounds anticancer activity of Pattern Recognition.doc

Probabilistic neural network for aroyl compounds anticancer activity of Pattern Recognition 【Abstract】 The Bayes classification based on statistical thinking and the rules of the basic network probabilistic neural network model prediction Fang acyl compounds anticancer activity. Choose 24 samples for training set, 6 sample prediction set, results show that the probabilistic neural network has good generalization ability, its ability to identify superior to Fisher discriminant method and fuzzy k-means clustering method. Keywords: aryl acyl compounds; probabilistic neural network; anti-tumor activity; pattern recognition Aryl acyl compounds can inhibit the nucleotide reductase activity, thereby inhibiting cancer cell growth. Elford et al [1] these compounds inhibit nucleotide was determined by the semi-reductase inhibition affected the amount of PC and for the L1210 tumors in mice by aryl acyl compounds, the average life expectancy after treatment with untreated mice, the percentage of the average life expectancy T / C, T / C is active in vivo parameters, T / C greater, then the stronger the anti-tumor activity. However, previous studies found that aromatic acyl compounds the electronic structure of drugs index T / C is not a good quantitative structure-activity relationship. Since the drug effect in vivo anti-tumor activity involving drugs reach the receptor and the role of drug-receptor complex process, affecting more factors to examine them more difficult, while the pattern recognition method does not require a precise mathematical model requires a priori knowledge of less [2], which study the anticancer activity of drugs provides a powerful tool. In this study, probabilistic neural network (probabilistic neural network, PNN) for the aryl acyl compounds anticancer activity of pattern recognition, with satisfactory results. Principle and a probabilistic neural network structure [3 ~ 5] PNN probabilistic neural network is a branch of the radial ba