Optimization Algorithm for English Vocabulary Learning Based on Long Short-Term Memory Networks

With the increasing importance of English vocabulary learning in language education, traditional vocabulary learning methods often face the problems of low learning efficiency and unstable memory retention. How to efficiently learn a large number of words in a short time and memorize them for a long time has become a challenge to be solved in the field of language learning. To this end, this paper proposes an English vocabulary learning optimization algorithm based on long shortterm memory network (LSTM), which aims to improve the efficiency of vocabulary learning and memory retention by combining deep learning technology with personalized learning strategies. This paper adopts LSTM as the core technical framework, combines adaptive learning rate and review mechanism based on forgetting curve, and designs an optimized vocabulary learning system. Through comparative experiments, this paper verifies the advantages of the algorithm in vocabulary memory accuracy, learning time and memory retention. Experimental results show that compared with traditional learning methods and other deep learning models (such as RNN, GRU), the LSTM model can significantly improve the efficiency of vocabulary learning, reduce learning time by $15 \%$, and increase long-term memory retention by $5 \%$.