Research on Risk Prediction Based on Machine Learning and Decision Making Framework

Based on the Topsis model, this study focuses on the challenges posed by extreme weather phenomena and highlights the application of the ARIMA time series model in risk forecasting. First, we used the Topsis model to correlate the indicators affecting the risk of extreme weather and, in conjunction with the gamma distribution, we analyzed the relationship between the local risk level and the relevant variables. Crucially, the ARIMA model is applied to time series forecasts of the number of insured people and the number of people affected by disasters, providing data-driven decision support for risk assessment. On this basis, we extended the model to an entropy-based Topsis model, which calculates the weights of different influencing factors by the entropy weight method to more accurately assess the construction needs and potential risks of the region. Meanwhile, combining the Delphi method and hierarchical analysis, we constructed the Topsis evaluation model to analyze the cultural value of the building and propose corresponding protection measures. Finally, based on the above model, the historical landmarks were systematically evaluated and targeted protection programs were proposed to enhance their disaster prevention capabilities. Overall, this study provides a complete set of machine learning and decision-making framework for risk management under extreme weather events, which has important application value, especially in the fields of urban planning and historical landmark protection.