The Application of Intelligent Recommendation Systems in Personalized Learning Path Planning: Combining CNN and Genetic Algorithms

Aiming at the problem that traditional personalized learning path planning methods rely on expert experience and are difficult to accurately capture the deep relationship between learners' needs and resources, convolutional neural networks and genetic algorithms are combined for personalized learning path planning. A deep CNN architecture is designed to extract features from learning behavior logs, and multi-layer convolution kernels are used to automatically capture learners' potential features, so as to accurately describe students' learning status. The feature vector is input into the GA module as the encoding basis of its initial population, and a chromosome encoding scheme containing learning resource sequences is designed. The tournament selection method is used to retain high-quality individuals. 20 students are selected for research, and the average recommendation matching degree of the CNN-GA model is 0.8495. The CNN-GA hybrid model effectively solves the problems of insufficient precision and poor adaptability in personalized learning path planning through the synergy of feature extraction and path optimization.