There is a copying effect in women's mate selection process, but there is no clear conclusion on the key factors that affect the copying effect. Thus, we investigated this issue based on the Classification and Regression Tree (CART) and Adaboost methods in machine learning. Firstly, a research scenario was constructed based on a well-known Chinese dating platform, where the dependent variable was set to determine whether the copying effect of highly educated female members was significant in the mate selection process. The independent variable selected several characteristics of women, such as age, emotional experience, income, and appearance. Then, based on the CART method and Adaboost method, a classification model for the copying effect of highly educated women's mate selection was constructed. The experimental results on the test dataset showed that the Fl-Score of the CART model was 82.4%, and the Fl-Score of the Adaboost model was 88.7%, both of which had high accuracy. Age, emotional experience, income, appearance, and other characteristics do play a classification role in the above classification model. These four variables were key factors in the copying effect of highly educated women's mate selection.