对偏最小二乘(PLS)回归的基本方法进行了分析研究,提出了基于非线性迭代偏最小二乘(NIPLS)的信息模式识别算法。该算法实现了模式识别中特征提取与分类器设计的有机结合。NIPLS较Fisher判别分析、Bayes判别分析等经典的模式识别算法,具有更强的信息识别能力,且对数据本身的分布要求不高,尤其对于多重共线性资料或解释变量多而样本数量少时更为有效。将该算法应用于土地质量的分类识别,结果表明,该文所建立的算法是有效的、可靠的。
On the basis of analysis and studies on modeling method of Partial Least squares (PLS), a novel algorithm of information pattern recognition based on Nonlinear Iterative Partial Least Squares (NIPLS) is set up in this paper. This algorithm reasonably combines feature extraction with classifier design, and has more advantages than classical algorithm of Fisher Discriminant Analysis (FDA) and Bayes Discriminant Analysis (BDA), such as simplicity and robustness, clearly qualitative explanation, and applies it to classification recognition of land quality. The simulation results show that the algorithm has better recognition effect than FDA and BDA, It is powerful for multicollinearity, particularly when the number of predictor variables is large and the sample size is small, and provides a novel and efficient algorithm for pattern recognition.