基于增强核极限学习机的专业选择智能系统

An intelligent system based on enhanced kernel extreme learning machine for choosing the second major

基于粒子群优化(PSO)的增强型核极限学习机(KELM)提出了一种有效的预测模型PSO-KELM来辅助第二专业选择。在PSO-KELM中,PSO策略确定KELM的最佳参数。PSO-KELM与其他两个竞争方法在学生专业选择数据上通过10折交叉验证方案进行比较,这两个方法分别是支持向量机和网格搜索技术优化的KELM。结果表明了本文预测模型在分类精度、受试者工作特征曲线面积(AUC)、灵敏度和特异性方面的优越性。

This paper proposes an effective prediction model for choosing the second major based on the Particle Swarm Optimization（PSO）enhanced Kernel Extreme Learning Machine（KELM）,which is called PSO-KELM model.In this model,the PSO strategy is adopted to adaptively determine the optimal parameters in KELM.The PSO-KELM model is compared with other two competitive methods,including Support Vector Machine（SVM）and a KELM is optimized by grid search technique,on a major selection dataset via a 10-fold cross validation scheme.The results clearly confirm the superiority of the proposed PSO-KELM model in classification accuracy,area under the receiver operating characteristic curve（AUC）,sensitivity and specificity.