基于特征优化的LSTM方法研究及其工业软测量建模应用.docxVIP

基于特征优化的LSTM方法研究及其工业软测量建模应用

关键词：特征优化；LSTM；工业软测量；深度学习；时间序列预测

Abstract:Withthecontinuousimprovementofindustrialautomationandinformatizationlevels,thereisanincreasingdemandforreal-timeandaccuratedatamonitoringandanalysis.TraditionaltimeseriespredictionmodelssuchasARIMAandSARIMAhavelimitationsinhandlingnonlinearandnon-stationarytimeseriesproblems.LongShort-TermMemorynetworks(LSTM)areatypeofadvanceddeeplearningmodelthatpossessesstrongfeaturelearningcapabilitiesandgoodtimeseriespredictionabilities.Thispaperaimstoexplorethemetho