Research on Oil Temperature Time Series Prediction Algorithm Based on Deep Learning

With the swift advancement of deep learning technology, time series prediction has become a hot topic in many research fields. However, traditional prediction models have limitations in dealing with complex nonlinear features and seeking global optimal solutions. With the aim of solving this problem, a model combining a Temporal Convolutional Network based on Sparrow Search Algorithm and a Bi-directional Long Short-term Memory Network (SSA-TCN-BiLSTM) is introduced. The model combines the advantages of two network structures to capture sequential patterns on different time scales. The experimental results highlight the model's outstanding prediction accuracy on various datasets, showcasing its ability to effectively learn and represent complex time-series patterns.