FOA优化极限学习机算法及模型应用研究

Research on FOA Optimization Extreme Learning Machine Algorithm and Model Application

传统极限学习机算法输入权重和单元偏置是随机确定的,这会导致算法性能不高,因而采用果蝇优化算法进行优化。分析了果蝇优化算法的流程,并和遗传算法、粒子群算法进行对比,通过求解Schaffer函数最优化问题验证了果蝇优化算法具有良好的优化性能。在此基础上,将果蝇优化算法优化极限学习机,提出改进的极限学习机算法。将改进极限学习机算法和极限学习机算法应用于化工企业竞争力预测中,结果表明,改进极限学习机算法对化工企业竞争力预测的准确率明显高于极限学习机算法,同时运行时间相差非常小。FOA优化极限学习机模型对其他预测问题的解决提供了参考。

The input weight and unit offset of traditional extreme learning machine algorithm are determined randomly,which leads to the bad performance of the algorithm,hence,the fruit fly optimization algorithm is adopted for improvement.The flow of fruit fly optimization algorithm(FOA)is analyzed.Three optimization algorithms,e.g.,fruit fly optimization algorithm,genetic algorithm and particle swarm optimization algorithm,are compared.FOA has good optimization performance by solving Schaffer function.Then,FOA is used to optimize extreme learning machine,and an improved extreme learning machine algorithm is proposed.The improved extreme learning machine algorithm and extreme learning machine algorithm are applied to the competitiveness prediction of chemical enterprises.The results show that the accuracy of the improved extreme learning machine algorithm for the competitiveness prediction of chemical enterprises is significantly higher than that of extreme learning machine algorithm,and the difference in running time is very small.The FOA optimized extreme learning machine model provides a reference for solving other prediction problems.