Abstract: Extreme Learning Machine（ELM） is not stable in predicting stock price. To address the problem, this paper proposes an Improved Fruit Fly of Algorithm（IFOA） that optimizes ELM by improving fruit fly. In the algorithm, fruit fly swarm constantly adjusts its radius to optimize the ELM input weights and thresholds of hidden layer, building an ELM prediction model with the optimized results. The IFOA-ELM model proposed in this paper can be applied to predicting stock price. Compared with ELM and FOA-ELM, IFOA-ELM model is more accurate and stable in predicting stock price.

Key words: stock price, prediction, Fruit Fly of Algorithm（FOA）, Extreme Learning Machine（ELM）

摘要： 针对股票价格预测中应用极限学习机预测存在稳定性不理想的问题，提出了一种改进果蝇优化极限学习机（IFOA-ELM）预测模型的算法。在该算法中，果蝇群通过不断调整群半径来优化ELM的输入层与隐含层连接权值和隐含层阈值，并以优化后的结果为基础，构建ELM预测模型。将IFOA-ELM模型用于股票价格预测。实验表明，与ELM和FOA-ELM相比，IFOA-ELM在股票价格预测中具有更高的预测精度和更好的稳定性。

关键词: 股票价格, 预测, 果蝇优化算法, 极限学习机