用 dataset3 作为训练数据，用 dataset4 作为测试数据，采用不同的特征、训练样本数、分类方法进行比较实验，观察、分析实验结果的异同。 训练分类器的方法为最小错误率贝叶斯分类器（假设正态分布，先验概率各 50%）。使用Bayesscr ipt.m运行代码。(Using dataset3 as training data, dataset4 is used as test data, and different characteristics, training samples and classification methods are used for comparative experiments, and the similarities and differences of experimental results are observed and analyzed. The method of training classifier is the minimum error rate Bias classifier (assuming normal distribution, the priori probabilities are 50%))