针对支持向量机的参数选择问题，本文提出了一种采用细菌群体趋药性智能优化算法优化最小二乘支持向量机参数的方法。细菌群体趋药性智能优化算法引入了群体信息交互策略，单个细菌不仅利用自身信息随机移动，而且细菌群体之间交换种群的信息，有效地改善了个体移动时的随机性和盲目性，加强了细菌趋于最优的移动策略。该方法提高了支持向量机的参数选择效率，避免了人为设定参数的不足，大大缩短了优化时间。经过细菌群体趋药性智能优化算法优化得到的最小二乘支持向量机的参数对，用于测试样本的多分类实验和函数拟合实验，其分类结果和函数拟合效果验证了本文方法的有效性。

A method that optimizes the parameters of the Least Squares Support Vector Machines (LS_SVM) by using Bacterial Colony Chemotaxis is presented.This optimization algorithm is used to optimize the target function by analogy to the way bacteria react to chemoattractants in concentration gradients. This method greatly improves the efficiency of LS_SVM’s parameters selection, and with the parameters selected, the classification result for the testing samples is the optimum. It avoids the disadvantage of manually specifying the parameters, and also scales down the optimisation time.