The Improved Bi-LSTM-Transformer Model and its Application

As a popular machine learning algorithm, deep learning can be applied to a wide range of fields. In this paper, the LSTM model in deep learning, its extended Bi-LSTM model, the Transformer model for processing sequence data and the Bi-LSTM-Transformer combination model are selected to conduct research and analysis, and the time series data under the influence of multiple factors are compared and studied, and the aging data are listed to compare and verify the models, and the experimental results are collected. According to the advantages of the above model in dealing with time series data, the parameters of the model are continuously optimized. The comparison of the evaluation index results shows that the combined model has higher accuracy than the other three models in predicting data, and the goodness of fit of the model reaches 0.994, and MAPE is only 0.565%. This result shows that the Bi-LSTM-Transformer combined model has better prediction ability in fitting time series data under the influence of multiple factors.