Projection Pursuit Method Based on Connection Cloud Model for Assessment of Debris Flow Disasters

Abstract

A rational evaluation of the danger of debris flow disasters at the regional scale is essential for developing effective disas-ter prevention measures and economic planning in debris flow-prone areas. A novel projection pursuit method based on the connection cloud model and fruit fly optimization algorithm is addressed to analyze the dangerous degree of debris flow disasters at the regional scale, considering the random and fuzzy uncertainties of the projection direction vector. In this method, the connection cloud model gen-erates the candidate projection directions around the latest optimization; these candidate projection direction vectors are screened based on set pair analysis to advance the convergence rate. Case studies and comparisons with other algorithms are further carried out to verify the validity and reliability of the proposed method. Results demonstrate that the proposed method does not require existing evaluation criteria compared to the conventional evaluation methods. It can describe the randomness and fuzziness of the projection direction vector and better find the structural characteristics of fuzzy indicators randomly distributed in the finite intervals with a quicker convergence rate.

Keywords: projection pursuit, fruit fly optimization algorithm, debris flow, connection cloud model, assessment