关键词: 采空区, 评价指标, 聚类算法, 信息熵 

Abstract: With the in-depth application of intelligent algorithms in the field of disaster evaluation,it is very important to construct a reasonable evaluation index system for the efficient evaluation of surface safety in complex goafs. However,the traditional index selection method has many bottlenecks,such as strong subjectivity,many interference factors,low efficiency and insufficient digitization. Therefore,an efficient and accurate evaluation index model based on K-Means clustering algorithm and IE theory is constructed. Firstly,the model starts from the mechanism of surface disasters in goaf and extensively screens potential evaluation indicators. Then,the K-Means algorithm is used to cluster and filter these indicators to reduce the redundancy and complexity of the indicator information expression. Through IE theoretical calculation,the key indicators that have a significant impact on safety are extracted,and a set of index system for safety evaluation of complex sites in goaf is constructed. In order to verify the rationality of the index system,PCA and entropy weight method are used to test and evaluate. The model is applied to a goaf area and compared with the evaluation results of common methods. The results show that the model successfully simplifies 38 primary indicators to 8 key indicators,and the constructed evaluation index system can represent 87. 9% of the original index information with only 21. 1% of the index characteristics,which significantly reduces the calculation workload and improves the evaluation efficiency. The research results not only provide a new method for the evaluation of surface stability in goaf,but also provide theoretical support for the research in related fields,which has high theoretical value and practical significance. 

Key words: goaf areas,evaluation indicators,clustering algorithms,information entropy