An Uncertain Information Fusion Method for Distribution Demand Prediction of Emergency Materials

Abstract

The invention discloses a method for predicting the distribution demand of emergency supplies based on the fusion of uncertain information. Firstly, the basic attributes of each example in the historical disaster database are analyzed, the common features of the examples are extracted, and the rough set method is used to simplify, and the supplies are predicted according to each condition attribute. Calculate the attribute weight value based on the degree of influence of the results; then use the K-means method of feedback compensation to classify the reduced disaster data, and judge which category the current case belongs to according to the distance between the current case and various center points. According to the calculation method of the case similarity, the historical target case with the highest similarity is retrieved; finally, the material demand of the current case is linearly estimated according to the material consumption of the target case and the weight value of the case attributes. In addition, the rescue priority of the current case can be derived according to the reduced paradigm. Through this method, the problem of forecasting the demand distribution of rescue supplies is solved, and it provides a basis for the dispatching decision of the aviation emergency rescue command system.

Description

technical field [0001] The invention relates to a forecasting technology for emergency material demand, in particular to a multi-source information fusion method and an information rule mining method for emergency events under the uncertainty of rescue information, which provides a basis for accurately implementing emergency material dispatch and rescue. Background technique [0002] In various measures of disaster relief and handling emergencies, obtaining accurate rescue information, especially rescue material demand information, is a prerequisite for emergency coordinated dispatching decisions. At present, the demand forecast of emergency relief materials is mainly researched from the following aspects: (1) The expert experience prediction method of emergency material demand mainly uses the data and related theories of historical earthquake disaster analysis, and according to the statistical magnitude, number of deaths, number of injured, collapsed Housing and other param...

AI Technical Summary

Problems solved by technology

[0003] However, the demand forecast for emergency rescue resources is affected by many social and environmental factors, and has a strong timeliness and phases. Limited, it is necessary to combine the forecast of resource demand distribution based on the mining of historical statistical data in the disaster area to ensure the accuracy of the forecast

Previous models lack the ability to combine historical material demand data mining rules into uncertain information fusion emergency material demand forecasts, so it is difficult to obtain accurate emergency rescue material demand

Images