Potentiality Evaluation for Revegetation of abandoned lands from coal mining activities based on Support Vector Machine

Activities of excavating coal exscind natural vegetation and deposit stone on the natural land that modified the natural land contribute to waste farmlands. Additionally, pollution of environment, losses of human life, human settlements and the infrastructure are also rising, which certainly demands urgent attention. Revegetation had been considered as a kind of cost-effective means. A reasonable potentiality for Revegetation of abandoned lands from coal mining activities is benefit for planning of Revegetation. In this paper, 34 instances were investigated, and seven attributes such as slope angle, elevation, topographic wetness index, lineaments, geological formations, soil types, condition of traffic, correlative with abandoned lands were recorded. A potentiality evaluation method based on support vector machine was proposed and was tested on those data. The purpose of SVM proposed in this paper is to construct a model that suggests target value of data instances in the testing set using only the given attributes. We randomly select 22 data instances to construct the SVM model. Testing is made by rest data. The results show SVM the good performance of potentiality evaluation with RBF kernel. it manages to achieve 95% success on the training set and 75% success on the testing set. Experiments performed also show that the performance of this method is mostly superior to that of artificial neural networks and SVM can be employed as an efficient method for evaluating the revegetation potentiality of abandoned lands from coal mining activities.