Due to the multiple composition of TCM (Traditional Chinese Medicine) data, SOC (Service Oriented Computing Quality) is difficult to be guaranteed when it is fused. Therefore, a deep fusion method of TCM multi-source heterogeneous big data based on knowledge atlas is proposed. Use crawler technology to obtain multi-source heterogeneous big data of traditional Chinese medicine and save it to the database. The knowledge map of traditional Chinese medicine was constructed by rule-based knowledge reasoning learning. In the data fusion stage, the random forest algorithm is introduced, taking the boundary of TCM knowledge data information classification as the benchmark, the generalization error of TCM knowledge data information is less than the data allowed to go online, and is fused into the corresponding TCM knowledge data tree to achieve the deep fusion of TCM multi-source heterogeneous big data. The test results show that when the pollution data increases from 10% to 30.0%, the ROC of the fused data decreases from 0.956 to 0.916 by only 0.040. This study provides a unified representation framework for these multi-source heterogeneous data by constructing a knowledge graph of traditional Chinese medicine, enabling the data to be more effectively integrated and utilized.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

A Knowledge Graph-Based Approach for Deep Fusion of Multi-source Heterogeneous Big Data in Chinese Medicine

  • Yao Fu,
  • Hao Wang

摘要

Due to the multiple composition of TCM (Traditional Chinese Medicine) data, SOC (Service Oriented Computing Quality) is difficult to be guaranteed when it is fused. Therefore, a deep fusion method of TCM multi-source heterogeneous big data based on knowledge atlas is proposed. Use crawler technology to obtain multi-source heterogeneous big data of traditional Chinese medicine and save it to the database. The knowledge map of traditional Chinese medicine was constructed by rule-based knowledge reasoning learning. In the data fusion stage, the random forest algorithm is introduced, taking the boundary of TCM knowledge data information classification as the benchmark, the generalization error of TCM knowledge data information is less than the data allowed to go online, and is fused into the corresponding TCM knowledge data tree to achieve the deep fusion of TCM multi-source heterogeneous big data. The test results show that when the pollution data increases from 10% to 30.0%, the ROC of the fused data decreases from 0.956 to 0.916 by only 0.040. This study provides a unified representation framework for these multi-source heterogeneous data by constructing a knowledge graph of traditional Chinese medicine, enabling the data to be more effectively integrated and utilized.