本文应用逐步判别分析方法对疟原虫血涂片细胞进行了分类研究。用169个红细胞(其中正常红细胞为130个,带疟原虫的红细胞为39个)作为训练集,192个红细胞(其中正常红细胞为157个,异常的为35个)作为考试集进行了统计分析,并通过实验调整判别阈值。对考试集的判别:假阴性率为11.4%,假阳性率为7.6%,结果较为理想。
An algorithm of Stepwise Discriminant Analysis is used in the classification of Plasmodium blood smear. 9 features of the red cell is extracted. 169 red cells (among them 130 cells are normal, 39 contain Plasmodium) are used as training set, 192 red cells (among them 157 cells are normal, 35 contain Plasmodium) as test set. We have done the statistical analyses and got good results. For test set, the false negative rate is 11.4%, the false positive rate is 7.6%.
丁岩,柴振明,陈传涓.逐步判别分析在疟原虫血涂片细胞分类中的应用[J].生物化学与生物物理进展,1990,17(2):121-125
复制生物化学与生物物理进展 ® 2025 版权所有 ICP:京ICP备05023138号-1 京公网安备 11010502031771号