The invention provides an early
pregnancy risk prediction method and
system based on multi-
source data and a medium. Firstly, basic information, physiological cycle data and living
habit data of a user are collected through a preset interface; preprocessing the
original data to generate standardized structure data; extracting age features, periodic regularity features and living
habit features to obtain a
feature vector set; then, sequentially comparing the age feature with a first threshold value to generate a
risk index, comparing the physiological cycle feature with a second threshold value to update the
risk index, and matching the living
habit feature with a third threshold value to obtain
feature matching information; integrating the
risk index and the matching information based on a weighted fusion
algorithm to generate a
pregnancy risk
score; and finally, generating a visual risk interface through a graphic rendering engine in combination with historical data, and outputting the visual risk interface. Therefore, comprehensive assessment and real-time monitoring of
pregnancy risks are realized, and the accuracy of health management and the effectiveness of decision support are improved.