医学信息2025，Vol.38Issue(8)：15-20,6.DOI:10.3969/j.issn.1006-1959.2025.08.003
基于深度学习的急性白血病流式细胞术检测报告文本资料自动分类研究
Automatic Classification of Text Data for Flow Cytometry Detection of Acute Leukemia Based on Deep Learning
摘要
Abstract
Objective To explore the classification effect of deep learning model on text data of flow cytometry report results.Methods Six deep learning models such as CNN and LSTM were used to analyze the text data of the results of the flow cytometry report,classify and predict the patients with acute leukemia,and finally evaluate the model by the comprehensive index F1 score.Results The precision,recall and F1 score of the CNN-BiLSTM mixed model were the best,which were 0.7422,0.7365 and 0.7361,respectively,and the F1 score of the model reached 70%in seven categories:normal humans,acute myeloid leukemia,acute B lymphoblastic leukemia,nucleated red blood cell abnormalities,neutrophil abnormalities,plasma cell abnormalities and monocytic abnormalities.Conclusion The mixed model has a good effect on the classification of text data in the results of flow cytometry test report,and can be combined with previous studies to build a more complete automated flow cytometry analysis system to further improve the efficiency and accuracy of flow cytometry analysis.关键词
流式细胞术/文本分类/CNN/自动化分析/深度学习Key words
Flow cytometry/Text classification/CNN/Automated analysis/Deep learning分类
信息技术与安全科学引用本文复制引用
张亚洲,李智伟,农卫霞,雷伟,摆文丽,李寅臻,李瑞,王奎..基于深度学习的急性白血病流式细胞术检测报告文本资料自动分类研究[J].医学信息,2025,38(8):15-20,6.基金项目
国家自然科学基金项目(编号:81860374) （编号:81860374）
