废旧电力电缆分级分类模型研究

Research on Classification Model of Waste Power Cables

摘要: 为提升废旧电缆处置效率与效益，提高国网上海市电力公司废旧物资的资产回收水平，本文通过建立废旧电缆评价体系，提取对废旧电缆价值影响显著的特征指标；通过二阶聚类法以提取出的特征指标为主要因素对历史数据进行分类；通过决策树模型对废旧电缆各型号分类结果进行预测；最后对决策树预测性能进行验证，并产生风险统计表。结果表明，预测正确率在90%以上，风险均小于0.05，可见分类结果优良。

Abstract: In order to improve the efficiency and benefit of waste cable disposal and improve the asset re-covery level of waste materials of State Grid Shanghai Electric Power Company, this paper extracts the characteristic indicators that have a significant impact on the value of waste cables by estab-lishing a waste cable evaluation system; through the second-order clustering method, the historical data is classified with the extracted feature index as the main factor; the classification results of various types of waste cables are predicted through the decision tree model; finally, the prediction performance of the decision tree is verified and a risk statistics table is generated. The results show that the prediction accuracy rate is above 90%, and the risk is less than 0.05, which shows that the classification results are excellent.