Skip to main content
Fig. 2 | BioPsychoSocial Medicine

Fig. 2

From: Association between maternal overprotection and premenstrual disorder: a machine learning based exploratory study

Fig. 2

Visualization of model performance and contributing features. a Model performance, based on the area under the receiver operating characteristic (ROC) curve. b Features with higher overall Shapley Additive exPlanations (SHAP) scores, indicating higher average contributions to premenstrual disorder (PMD) prediction. c Positive and negative contributions to PMD, characterized by positive and negative SHAP values, respectively. Red and blue represent the magnitude of the values. AUC, area under the receiver operating characteristic curve; PBIm/f, Parental Bonding Instrument (mother/father); MDQ, Menstrual Distress Questionnaire; PSQI, Pittsburgh Sleep Quality Index; BP, blood pressure (s, systolic; d, diastolic; m, mean); SOC, Sense of Coherence; STAI, State–Trait Anxiety Inventory. Items PBIm_23, PBIm_10, and PBIf_23 are scored as follows: very like, 3; moderately like, 2; moderately unlike, 1; and very unlike, 0. Items PBIm_2 and PBIm_15 were reverse scored. The PBIm_care and PBIf_care scores range from to 0–36. The PBIm_overprotection scores range from to 0–39. For information on the other measures, see Additional file 1

Back to article page