Intelligent recommendation of educational resources combining Neu-MF and T-S fuzzy control

Abstract

The research uses Takagi-Sugeno (T-S) fuzzy control combined with neural matrix factorization (Neu MF) model to study the intelligent recommendation of educational resources. The recommendation performance of TS-Neu MF model is compared with other similar recommendation algorithm models under two test sets of E's dx and C er. The results of the experiments show that the TS-Neu MF model outperforms Deep FM by 56.6% in root mean square error (RMSE) metrics and 71.5% in mean absolute error (MAE) metrics, and outperforms the Neu MF model by 33.1% in RMSE metrics and 22.5% in MAE metrics under the E dx dataset. The training loss is about 0.04 lower than the Deep FM model, about 0.006 lower than the BPNN model, and about 0.02 lower than the Neu MF model.