Automated acquisition of bone tumor diagnostic knowledge by data mining technology

    ZHANG Hui, LI Jun, QIAN Zongª²Cai, QU Jingª²Hui, FAN Qingª²Yu

    1Department of Computer Application, School of Biomedical Engineering, Fourth Military Medical University, Xi¡¯an 710033, China, 2Institute of Bone Tumors of Chinese PLA, Tangdu Hospital, Fourth Military Medical University, Xi¡¯an 710038, China

    ¡¾Abstract¡¿ AIM: To solve the bottleneck problem in establishing computerª²aided bone tumor diagnostic expert system and to find a way to acquire bone tumors diagnostic knowledge automatically and set up knowledge database from diagnosed cases. METHODS: Data mining technology based on the rough set theory was used to pretreat the diagnosed cases and data reduction, to acquire bone tumor diagnostic knowledge automatically and set up diagnostic knowledge base. RESULTS: Program module of knowledge automated acquisition created a decisionª²making tree with 29 nodules, which reduced the condition attributes in database from 43 points to 28 points, including 6 decisionª²making attributes. It significantly improved the operation speed of the subsequent discursion programs CONCLUSION: Automated acquisition by data mining technology can help to set up diagnostic knowledge database and to accelerate the establishment of expert system with higher reliability and accuracy. ......