基于ISSA-CNN-LSTM模型对青岛市的平均温度预测
A Prediction of Average Temperature in Qingdao Based on ISSA-CNN-LSTM Model

针对于目前对平均温度预测精度不高的问题，提出了一种ISSA-CNN-LSTM模型。针对于传统的麻雀搜索算法(SSA)的缺点而言，提出了一种改进的SSA模型(ISSA)，通过引入Cat映射初始化、Tent扰动和柯西扰动等技术，有效避免了初始解过于集中带来的局部最优问题。此外，Tent扰动和柯西扰动方法在粒子更新过程中引入了更强的扰动，增强了搜索的跳跃性和广度，避免了传统SSA模型中对局部解的过度依赖。其中CNN为卷积神经网络，LSTM为长短期记忆网络。仿真结果表明，此模型对于相对湿度的预测优于其他模型。
In response to the low prediction accuracy of average temperature at present, an ISSA-CNN-LSTM model is proposed. In response to the shortcomings of the traditional sparrow search algorithm (SSA), an improved SSA model (ISSA) is proposed by introducing Cat mapping initialization, Tent perturbation, and Cauchy perturbation techniques, effectively avoiding the local optimal problem caused by the initial solution being too concentrated. In addition, the Tent perturbation and Cauchy perturbation methods introduce stronger perturbation in the particle update process, enhance the search’s jumping and breadth, and avoid the excessive dependence on local solutions in traditional SSA models. Among them, CNN is a convolutional neural network, and LSTM is a long short-term memory network. Simulation results show that this model has better prediction accuracy for relative humidity than other models.