邰杨芳, 昝彭, 华国旻. 基于疾病网络的血友病并发症挖掘与关联规则分析[J]. 实用临床医药杂志, 2024, 28(3): 6-12. DOI: 10.7619/jcmp.20233592
引用本文: 邰杨芳, 昝彭, 华国旻. 基于疾病网络的血友病并发症挖掘与关联规则分析[J]. 实用临床医药杂志, 2024, 28(3): 6-12. DOI: 10.7619/jcmp.20233592
TAI Yangfang, ZAN Peng, HUA Guomin. Mining of hemophilia complications based on disease network and association rule analysis[J]. Journal of Clinical Medicine in Practice, 2024, 28(3): 6-12. DOI: 10.7619/jcmp.20233592
Citation: TAI Yangfang, ZAN Peng, HUA Guomin. Mining of hemophilia complications based on disease network and association rule analysis[J]. Journal of Clinical Medicine in Practice, 2024, 28(3): 6-12. DOI: 10.7619/jcmp.20233592

基于疾病网络的血友病并发症挖掘与关联规则分析

Mining of hemophilia complications based on disease network and association rule analysis

  • 摘要:
    目的 基于疾病网络探讨血友病并发症的一般性规律,预测血友病患者可能发生的并发症。
    方法 从PubMed数据库中检索血友病相关文献,通过MetaMap工具从文献标题、摘要文本中抽取疾病实体,基于疾病对的共现关系构建疾病网络,分析疾病网络的整体特征、节点特征及结构特征等。对疾病实体网络进行关联分析,挖掘其关联规则,分析血友病并发症的一般性规律。采用链路预测算法预测血友病的潜在并发症。
    结果 血友病及其并发症构成的关联网络在网络结构上满足小世界网络特征和分布均匀的凝聚子群特征。凝聚子群分析结果显示,血友病并发症可分为遗传性疾病、血液系统疾病、传染性疾病和慢性疾病共4大类。关联规则分析发现133条置信度≥0.8的规则,链路预测进一步得到了许多有据可查的疾病对。
    结论 基于疾病网络进行血友病并发症关联分析和链路预测,可实现对血友病潜在并发症的有效预测,为血友病的临床诊疗提供决策支持。

     

    Abstract:
    Objective To explore the general rules of complications of hemophilia based on the disease network and predict the possible complications of hemophilia patients.
    Methods The PubMed database was searched for literature related to hemophilia. Disease entities were extracted from the titles and abstracts of the literature using the MetaMap tool. A disease network was constructed based on the co-occurrence relationship of disease pairs, and the overall characteristics, node characteristics, and structural characteristics of the disease network were analyzed. The disease entity network was analyzed for association rules to explore the general patterns of hemophilia complications. The link prediction algorithm was used to predict potential complications of hemophilia.
    Results The association network composed of hemophilia and its complications satisfied the characteristics of small-world networks and evenly distributed cohesive subgroups in terms of network structure. Cohesive subgroup analysis showed that hemophilia complications can be divided into four categories: inherited diseases, blood system diseases, infectious diseases, and chronic diseases. Association rule analysis identified 133 rules with a confidence level ≥0.8, and link prediction further identified many well-documented disease pairs.
    Conclusion Based on the disease network, hemophilia complication association analysis and link prediction can effectively predict potential hemophilia complications, providing decision support for clinical diagnosis and treatment of hemophilia.

     

/

返回文章
返回