TY - GEN
T1 - Expert Data - Assisted Diagnosis
T2 - 18th International Conference on Knowledge Science, Engineering and Management, KSEM 2025
AU - Liu, Runze
AU - Zhang, Yingying
AU - Sheng, Hao
AU - Wang, Jingyi
AU - Gu, Xiaoyan
AU - Han, Jiancheng
AU - Yang, Da
AU - Huang, Xuefei
AU - He, Yihua
AU - Zhu, Haogang
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2026.
PY - 2026
Y1 - 2026
N2 - In prenatal screening for fetal congenital heart disease (CHD), ultrasonic diagnosis and other methods are prone to being affected by regional resource differences and insufficient experience in diagnosing doctors, thus resulting in misdiagnosis of cases. This research puts forward a combined discriminative system, which integrates the iTransformer method and XGBoost to aid in the prenatal diagnosis of fetal CHD. This system, named INFO-iTransformer-XGBoost, merges a combined discriminative system, INFO (Weighted mean of vectors optimization algorithm), and SHAP (SHapley Additive exPlanations) explainable analysis prediction model. By comparing the model results with those from INFO-iTransformer and INFO-XGBoost alone, the study confirms the advantage of the combined discriminative system in prenatal CHD screening for fetuses. The study used the fetal CHD detection dataset provided by the Maternal and Fetal Medicine Center of Beijing Anzhen Hospital, Capital Medical University, from February 2018 to August 2024. The research shows that the INFO-iTransformer-XGBoost combined discriminative system and SHAP model explainability analysis can provide a quantitative diagnosis and clinically interpretable diagnostic solution for prenatal CHD screening.
AB - In prenatal screening for fetal congenital heart disease (CHD), ultrasonic diagnosis and other methods are prone to being affected by regional resource differences and insufficient experience in diagnosing doctors, thus resulting in misdiagnosis of cases. This research puts forward a combined discriminative system, which integrates the iTransformer method and XGBoost to aid in the prenatal diagnosis of fetal CHD. This system, named INFO-iTransformer-XGBoost, merges a combined discriminative system, INFO (Weighted mean of vectors optimization algorithm), and SHAP (SHapley Additive exPlanations) explainable analysis prediction model. By comparing the model results with those from INFO-iTransformer and INFO-XGBoost alone, the study confirms the advantage of the combined discriminative system in prenatal CHD screening for fetuses. The study used the fetal CHD detection dataset provided by the Maternal and Fetal Medicine Center of Beijing Anzhen Hospital, Capital Medical University, from February 2018 to August 2024. The research shows that the INFO-iTransformer-XGBoost combined discriminative system and SHAP model explainability analysis can provide a quantitative diagnosis and clinically interpretable diagnostic solution for prenatal CHD screening.
KW - Combined Discriminative System
KW - Fetal congenital heart disease prediction
KW - SHAP explainability analysis
UR - https://www.scopus.com/pages/publications/105022976294
U2 - 10.1007/978-981-95-3058-8_26
DO - 10.1007/978-981-95-3058-8_26
M3 - 会议稿件
AN - SCOPUS:105022976294
SN - 9789819530571
T3 - Lecture Notes in Computer Science
SP - 291
EP - 298
BT - Knowledge Science, Engineering and Management - 18th International Conference, KSEM 2025, Proceedings
A2 - Zhu, Tianqing
A2 - Zhou, Wanlei
A2 - Zhu, Congcong
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 4 August 2025 through 7 August 2025
ER -