Research Paper Advance Articles

Machine-intelligence for developing a potent signature to predict ovarian response to tailor assisted reproduction technology

Schematic workflow for poor ovarian response prediction. (A) 11 features relating to ovarian response were obtained following logistic regression and LASSO. (B) These 11 candidate features were analyzed using multivariable logistic regression and machine learning, and then validated using ROC, calibration plot, C-index and correlation analysis to conduct CPLM and HPTM. (C) Variable importance of CPLM and HPTM were described to further understand and investigate of the models.

Figure 1. Schematic workflow for poor ovarian response prediction. (A) 11 features relating to ovarian response were obtained following logistic regression and LASSO. (B) These 11 candidate features were analyzed using multivariable logistic regression and machine learning, and then validated using ROC, calibration plot, C-index and correlation analysis to conduct CPLM and HPTM. (C) Variable importance of CPLM and HPTM were described to further understand and investigate of the models.