摘要
目前在多目标综合评价方面有诸多方法,层次分析法、模糊综合评价等对权重的确定主观性较大或计算过程过于复杂。本文在介绍支持向量机的基础上,建立ERP实施绩效评价指标体系,依据SVM算法建立评价模型,并与依据BP神经网络建立的评价模型进行对比,实验结果表明基于SVM的ERP实施绩效评价模型在小样本、数据灰度大的情况下更为高效和优越。这为ERP实施绩效评价问题提供了一种新的思路。
There are many ways to solve comprehensive evaluation of multi-objective problem, but the analytic hier- archy process, fuzzy comprehensive evaluation on determining the weights of subjectivity or calculation process is too complex. The paper firstly introduces the basic thought and steps of support vector machine, and constructs a comprehensive evaluation index system of ERP performance, then creates SVM based model and compared with BP neural network based model. Results are found that SVM based model is more efficient than BP Neural Network based model. It proved the SVM can work well and is a new way for ERP performance evaluation.
出处
《科技与管理》
2013年第5期63-66,共4页
Science-Technology and Management
