The invention discloses a
drug abortion success rate
prediction system and method based on a
random forest, particularly relates to the technical field of medical
data analysis, and aims to solve the problem that clinical doctors depend on experience to judge the success rate of
drug abortion and lack scientific and quantitative prediction tools. The method comprises the following steps: firstly, selecting a
target population which requires
drug abortion and has no
contraindication, and collecting clinical data of the
target population; making a binarization coding rule, coding the collected clinical data of the
target population item by item, and constructing a
feature vector of the target
population; based on the
feature vector of the target
population, predicting whether the drug abortion of the patient succeeds or not through a
random forest model; then collecting the drug abortion outcome of the target
population; and finally, merging the drug abortion outcome and the clinical data-
data model of the patient to obtain a predicted value of the drug abortion outcome of the patient. According to the invention, the
evaluation system for the adaptive population of the drug-induced abortion is established, the adaptive population of the drug-induced abortion is defined, and the
failure rate of the drug-induced abortion is reduced.