Personalized Learning Effectiveness Evaluation and Content Reorganization Based on Cognitive Diagnosis Theory and Machine Learning

With the rapid development of information technology, personalized learning has gradually become an important means of improving learning effects in education. At present, most personalized learning systems still face challenges such as inaccurate assessment of knowledge point mastery and insufficient personalization of recommended content. To solve these problems, this study proposes a personalized learning effectiveness assessment and content restructuring method based on cognitive diagnosis theory (DINA model) and collaborative filtering algorithms in machine learning. The method accurately assesses students' mastery of each knowledge item and recommends appropriate learning content based on similarities among students, thus improving the effectiveness of personalized learning. In this study, the DINA model is used to accurately assess students' mastery of knowledge items, and then the collaborative filtering algorithm is used to analyse students' learning behaviour and recommend personalized learning content. Through experiments on the learning data of 300 college students in advanced mathematics courses, the experimental results show that the hybrid method based on the DINA model and collaborative filtering has significant advantages over traditional recommendation methods in terms of learning content matching, accuracy of knowledge point mastery assessment, and learning progress. In particular, the recommendation system can effectively provide students with more targeted learning resources and help them to significantly improve their knowledge point mastery rate in a short period of time.