摘要
股价通常会受到非线性、时效性等因素的影响,传统的预测方法存在预测精度较低、方法复杂等问题,本文提出基于LSTM-Adaboost混合网络模型。对选取的股市历史数据进行平稳性处理,使用Adaboost集成学习训练出最优参数,得到最优预测模型,最后生成预测结果。结果表明,使用LSTM-Adaboost模型预测命中率提升了7个百分点,提升了股价预测的准确性。
The stock price is usually affected by non-linear,timeliness,and other factors.A hybrid network model is proposed in this paper.A stationary treatment was conducted on the historical data selected from stock market,and optimal parameters were obtained through training in Adaboost integrated learning,and an optimal prediction model was obtained to generate prediction results.The prediction hit ratio of the LSTM-Adaboost model increases by 7%,which improves the accuracy in predicting stock price.
作者
宁昱博
张玉军
NING Yubo;ZHANG Yujun(School of Computer and Software Engineering,University of Science and Technology Liaoning,Anshan 114051,China)
出处
《辽宁科技大学学报》
CAS
2019年第5期383-388,共6页
Journal of University of Science and Technology Liaoning