温度是影响危险废物贮存设施环境安全的关键指标,准确预测室内温度变化并采取相应防护措施可有效预防环境安全事故的发生。提出一种基于改进的双向长短期记忆(Bidirectional Long Short-Term Memory,BiLSTM)网络模型,通过互信息法和相...温度是影响危险废物贮存设施环境安全的关键指标,准确预测室内温度变化并采取相应防护措施可有效预防环境安全事故的发生。提出一种基于改进的双向长短期记忆(Bidirectional Long Short-Term Memory,BiLSTM)网络模型,通过互信息法和相关系数法筛选影响环境温度变化的关键特征,借助卷积神经网络(Convolutional Neural Network,CNN)提取数据中的局部特征,再利用双向长短期记忆网络捕获序列数据的前后依赖关系,并对危险废物贮存设施环境温度进行预测试验。结果表明,改进后的BiLSTM模型较为准确地预测了危险废物贮存设施未来一段时间的环境温度变化趋势,模型平均绝对误差(MAE)为0.166,平均绝对百分比误差(MAPE)为0.060,均方根误差(RMSE)为0.246,R^(2)为0.939,能够为危险废物贮存温度管控提供决策依据。展开更多
The strong approximations of a class of R^d-valued martingales are considered.The conditions usedin this paper are easier to check than those used in [3] and [9].As an application,the strong approximation ofa class of...The strong approximations of a class of R^d-valued martingales are considered.The conditions usedin this paper are easier to check than those used in [3] and [9].As an application,the strong approximation ofa class of non-homogenous Markov chains is established,and the asymptotic properties are established for themulti-treatment Markov chain adaptive designs in clinical trials.展开更多
基金Supported by The National Natural Science Foundation of China (No.10071072)
文摘The strong approximations of a class of R^d-valued martingales are considered.The conditions usedin this paper are easier to check than those used in [3] and [9].As an application,the strong approximation ofa class of non-homogenous Markov chains is established,and the asymptotic properties are established for themulti-treatment Markov chain adaptive designs in clinical trials.
