In the current landscape of digital transformation, objectively measuring and fostering students’ innovative and entrepreneurial competencies has emerged as a central mission of the educational community. Aiming at the defects of extreme subjectivity and lack of dynamics of the existing assessment methods, this paper constructs a comprehensive assessment model integrating AHP hierarchical analysis and BP neural network, and introduces the SHAP interpretation framework to reveal its decision-making mechanism. First, the multidimensional innovation and entrepreneurship ability evaluation indexes of students are established through feature engineering, and their weights are determined using AHP. The pre-processed samples are then fed into the BP neural network to obtain an accurate measure and evaluation of students’ innovation and entrepreneurship potential. SHAP post hoc interpretation is used to uncover the role played by each assessment index. Experimental data demonstrate the results show that the average prediction error between model outputs and measured values of the model in this paper is as small as 0.815% which is an indication of high precision. It also confirms that the innovation and entrepreneurial competence is among the major dimensions that determine the level of students’ overall innovation and entrepreneurship proficiency. It implements the dynamism behind individual innovation and entrepreneurial abilities of each student, thereby supporting the smooth operation of the learners’ careers.
