摘要
选取代表房地产上市公司综合实力的投资与收益、偿债能力、经营能力、资本结构等四个方面的10项指标,96个公司的财务数据,采用TOPSIS方法计算每个公司的综合绩效评价值,随机挑选其中的80组数据作为训练样本,16组数据作为测试样本,建立SVM模型,通过测试分析并与RBF神经网络预测模型的结果对比,表明SVM模型更加有效,更有推广前景。
The paper selects 10 indicators and financial data of 96 companies representing comprehensive listed real estate company strength fromfour aspects of investment and incom e, debt-paying a b ility , operation capacity and capital structure, calculates their own comprehensive performanceevaluation value by applying TOPSIS method, randomly selects 80 sets of data as training samples and 16 sets of data as testing samples, andfin a lly establishes SVM model. Through testing analysis and comparing to the results of RBF nerve network forecast m odel, it demonstrates that :SVM model is more effective and has wider promotion prospect.
出处
《山西建筑》
2016年第14期237-239,共3页
Shanxi Architecture
关键词
绩效评价
房地产
支持向量机
TOPSIS方法
performance evaluation, real estate, Support Vector Machine (S V M ) , TOPSIS method