Comprehensive Evaluation of Flood Disaster in Heilongjiang Province of China Based on RAGAPPC Model

摘要

Dimensionality reduction technology of high dimension data, Projection Pursuit Classification (PPC) Model, was used to evaluate flood disaster. In the case of flood disaster loss without evaluating criterion or with inapplicable criterion, Real-coded Accelerating Genetic Algorithm (RAGA) was utilized to optimize the projection direction of PPC model, and then transform multidimension data to low dimension subspace. Each index of flood disaster loss was evaluated based on seeking optimal projection direction and projection value, which avoided man-made disturbance of subjective weight determination and gained good results.