Indexed by:
Abstract:
本发明涉及一种面向水质指标预测的方法,特别是涉及一种基于SG(Savitzky‑Golay)滤波、变分模态分解(Variational Mode Decomposition, VMD)、注意力机制、基于遗传模拟退火的粒子群优化算法(Genetic Simulated annealing‑based PSO, GSPSO)与编码‑解码器的双向长短时记忆(Encoder‑Decoder Bidirectional Long Short Term Memory, BiLSTM‑ED)神经网络的水质指标预测方法。首先,将获取到的水质指标历史数据进行归一化处理,并采用SG滤波进行平滑去噪,其次对处理过后的数据使用VMD进行分解得到相对平稳的子序列,然后将数据输入加入注意力机制的基于编码‑解码器的双向长短时记忆BiLSTM‑ED神经网络模型,最后采用GSPSO优化模型的超参数,从而预测未来多个时间点的水质指标情况,最终获得了较高的预测精度。
Keyword:
Reprint Author's Address:
Email:
Patent Info :
Type: 发明申请
Patent No.: CN202211596666.9
Filing Date: 2022-12-12
Publication Date: 2023-03-31
Pub. No.: CN115879375A
Applicants: 北京工业大学
Legal Status: 实质审查
Cited Count:
WoS CC Cited Count: 0
SCOPUS Cited Count:
ESI Highly Cited Papers on the List: 0 Unfold All
WanFang Cited Count:
Chinese Cited Count:
30 Days PV: 0
Affiliated Colleges: