Innovation and entrepreneurship education is the key pivot for the in-depth implementation of “mass entrepreneurship and innovation”. The study goes beyond the single causal factor of innovation and entrepreneurship quality improvement, and introduces the fuzzy set qualitative comparative analysis (fsQCA) method, based on the questionnaires of 524 college students, to examine the combination paths of six dimensions: theoretical teaching, faculty allocation, practical activities, service platform, incentive policies and cultural atmosphere from a group perspective. The innovation and entrepreneurship education data are further utilized to portray individual student characteristics. Seventeen multi-dimensional behavioral data including online/offline learning, innovation practice, and output are obtained, and a fine-grained learner profile is constructed, and a deep belief network (DBN) model is constructed to realize the recommendation of innovation and entrepreneurship education curriculum resources. The study reveals five combination paths of “high faculty-practice-driven”, paths and “teaching-motivation-practice-driven” that lead to high activity in creative industries, with a consistency of 0.869. Based on students’ final grades and extracurricular practical activity scores, the study clustered five groups, namely, Overall Leaders (78), Academically Focused (146), Balanced Developers (169), Practice-Oriented (63), and Lagging Developers (69), to provide the coordinates for accurate profiling and interventions.The DBN achieved F1=0.34% and 95.19% for both student profile similarity matching and course prediction and recommendation. 91.34% and 95.19% excellent results.
