2019年人工智能技术分析报告.docx

2019年人工智能技术分析报告  正文目录 本文研究导读 ....................................................................................................................4 时序交叉验证的改进 .........................................................................................................5 K 折和时序交叉验证..................................................................................................5 改进思路 1——更合理的基线模型.............................................................................6 改进思路 2——更精细的切分方法.............................................................................7 方法...................................................................................................................................8 人工智能选股模型测试流程.......................................................................................8 单因子测试..............................................................................................................10 回归法和 IC 值分析法......................................................................................10 分层回测法 ......................................................................................................10 结果.................................................................................................................................11 最优超参数..............................................................................................................11 模型性能..................................................................................................................12 单因子测试..............................................................................................................13 构建策略组合及回测分析 ........................................................................................15 总结和讨论......................................................................................................................18 附录：分组时序交叉验证的代码实现..............................................................................19 修改 model_selection 包的_split.py ....................