%0 Journal Article
%T 基于随机森林与ARIMA模型的降水变化与灾害风险评估<br>Precipitation Variation and Disaster Risk Assessment Based on Random Forest and ARIMA Models
%A 崔培琪
%J Pure Mathematics
%P 237-258
%@ 2160-7605
%D 2025
%I Hans Publishing
%R 10.12677/pm.2025.151027
%X 随着对地观测技术的飞速发展&#65292;我们能够以前所未有的精度和频率获取地球表面的各种数据&#65292;通过进行更精细的空间分析和时间序列分析&#65292;可以揭示地理环境变化的深层次规律。本文旨在建立中国降水量变化趋势及其与海拔、坡度、土地利用之间的预测模型。通过对降水量、地形因素和五种主要土地覆盖类型的相关性分析&#65292;运用Logistic回归和随机森林模型探讨了这些因素对灾害发生的共同影响机制。此外&#65292;采用ARIMA时间序列模型预测了未来2025年到2035年间的降水量和土地利用格局&#65292;并结合随机森林模型评估了此期间各地区暴雨灾害风险的空间分布。研究结果揭示了在极端天气条件下最脆弱的地区&#65292;为灾害防范和土地规划提供了重要参考。<br>With the rapid advancement of remote sensing technologies, we are now able to obtain various data on the Earth&#8217;s surface with unprecedented accuracy and frequency. Through more refined spatial and time series analyses, the underlying patterns of geographical environmental changes can be revealed. This study aims to establish predictive models for the trends in precipitation changes in China and their relationships with elevation, slope, and land use. By analyzing the correlations between precipitation, topographic factors, and five major land cover types, the study employs Logistic regression and Random Forest models to explore the joint impact of these factors on the occurrence of disasters. Additionally, the ARIMA time series model is utilized to forecast precipitation and land use patterns from 2025 to 2035, while the Random Forest model is applied to assess the spatial distribution of rainfall disaster risks during this period. The results of the study highlight the most vulnerable regions under extreme weather conditions, providing valuable insights for disaster prevention and land planning.
%K Logistic回归模型&#65292
%K 随机森林预测模型&#65292
%K ARIMA时间序列模型&#65292
%K 中国土地利用变化&#65292
%K 暴雨成灾<br>Logistic Regression Model
%K Random Forest Prediction Model
%K ARIMA Time Series Model
%K Land Use Change in China
%K Rainstorm-Induced Disasters
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=106599