Li Jun, Yuan Hailong, Yu Xia, Hu Tian
International College of Culture and Tourism, Jilin International Studies University, Changchun, 130117, Jilin, China.
Qilu Institute of Technology, Jinan, 250200, Shandong, China.
Sci Rep. 2024 Jul 28;14(1):17342. doi: 10.1038/s41598-024-68457-w.
In order to augment the efficacy of the intelligent evaluation model for assessing the suitability of ice and snow tourism, this study refines the model by incorporating the Long Short-Term Memory (LSTM) network within the framework of the Internet of Things (IoT). The investigation commences with an elucidation of the application of IoT technology in environmental detection. After this, an analysis is conducted on the structure of LSTM and its merits in the realm of time series prediction. Ultimately, a novel model for appraising the suitability of ice and snow tourism is formulated. The efficacy of this model is substantiated through empirical experiments. The results of these experiments reveal that the refined model exhibits exceptional performance across diverse climatic conditions, encompassing mild, cold, humid, and arid climates. In regions characterized by mild climates, the predictive accuracy of the refined model progressively ascends from 88% in the initial quarter to 94% in the fourth quarter, surpassing the capabilities of conventional models. Consistently robust performance is demonstrated by the refined model throughout each quarter. In terms of operational efficiency, comparative analysis indicates that the refined model attains a moderate level, manifesting a 30-33 s runtime and maintaining a Central Processing Unit (CPU) usage rate between 40 and 43%. This observation implies that the refined model adeptly balances precision against resource consumption. Consequently, this study holds significance as a scholarly reference for the integration of IoT and LSTM networks in the domain of tourism evaluation.
为了提高冰雪旅游适宜性智能评估模型的效能，本研究在物联网（IoT）框架内纳入长短期记忆（LSTM）网络对模型进行优化。研究首先阐述了物联网技术在环境检测中的应用。在此之后，对LSTM的结构及其在时间序列预测领域的优点进行了分析。最终，构建了一个评估冰雪旅游适宜性的新型模型。通过实证实验验证了该模型的效能。这些实验结果表明，优化后的模型在包括温和、寒冷、潮湿和干旱气候在内的各种气候条件下均表现出色。在气候温和的地区，优化后模型的预测准确率从第一季度的88%逐步提高到第四季度的94% , 超过了传统模型的能力。优化后的模型在每个季度都表现出持续强劲的性能。在运行效率方面，对比分析表明，优化后的模型达到了中等水平，运行时间为30 - 33秒，中央处理器（CPU）使用率维持在40%至43%之间。这一观察结果表明，优化后的模型在精度与资源消耗之间实现了良好平衡。因此，本研究对于物联网和LSTM网络在旅游评估领域的整合具有学术参考意义。
