Core, well logging and seismic data were used to investigate sandbody architectural characteristics within Lower Member of Minghuazhen Formation in Neogene, Bohai BZ25 Oilfield, and to analyze the sedimentary microfac...Core, well logging and seismic data were used to investigate sandbody architectural characteristics within Lower Member of Minghuazhen Formation in Neogene, Bohai BZ25 Oilfield, and to analyze the sedimentary microfacies, distribution and internal architecture characteristics of the bar finger within shoal water delta front. The branched sand body within shoal water delta front is the bar finger, consisting of the mouth bar, distributary channel over bar, and levee. The distributary channel cuts through the mouth bar, and the thin levee covers the mouth bar which is located at both sides of distributary channel. The bar finger is commonly sinuous and its sinuosity increases basinward. The distributary channel changes from deeply incising the mouth bar to shallowly incising top of the mouth bar.The aspect ratio ranges from 25 to 50 and there is a double logarithmic linear positive relationship between the width and thickness for the bar finger, which is controlled by base-level changing in study area. For the bar finger, injection and production in the same distributary channel should be avoided during water flooding development. In addition, middle–upper distributary channel and undrilled mouth bar are focus of tapping remaining oil.展开更多
股票市场的准确预测对投资者和监管层而言都具有重要意义.在金融时间序列预测研究中分解-集成框架被广泛使用,然而以往研究中针对分解后的模态分量多数采用单一历史数据预测未来的思路,往往忽略了其他低频异质数据对分量的影响.本文融...股票市场的准确预测对投资者和监管层而言都具有重要意义.在金融时间序列预测研究中分解-集成框架被广泛使用,然而以往研究中针对分解后的模态分量多数采用单一历史数据预测未来的思路,往往忽略了其他低频异质数据对分量的影响.本文融合了分解-集成与混频数据采样思想,提出EEMD-Mixed Frequency CNN-BiLSTM-Attention/LSTM-LSTM(EE-MFCBA/L-L)股指收益率预测模型,通过EEMD将股指收益率分解为若干不同频率特征的分量,采用模糊熵算法识别分量频率特征,进而结合不同频率倍差的低频数据,使用MFCBA/L模型实现对模态分量的预测,最后采用LSTM模型非线性集成各分量的预测结果.实证结果表明,所提出的模型可以更好地适应收益率变化特征,与传统模型相比,所提模型在预测非平稳和非线性收益率序列方面具有显著优势,具有最低预测误差和最高的方向性预测准确率.展开更多
基金Supported by the National Natural Science Foundation of China(41772101)China National Science and Technology Major Project(2017ZX05009001-002)
文摘Core, well logging and seismic data were used to investigate sandbody architectural characteristics within Lower Member of Minghuazhen Formation in Neogene, Bohai BZ25 Oilfield, and to analyze the sedimentary microfacies, distribution and internal architecture characteristics of the bar finger within shoal water delta front. The branched sand body within shoal water delta front is the bar finger, consisting of the mouth bar, distributary channel over bar, and levee. The distributary channel cuts through the mouth bar, and the thin levee covers the mouth bar which is located at both sides of distributary channel. The bar finger is commonly sinuous and its sinuosity increases basinward. The distributary channel changes from deeply incising the mouth bar to shallowly incising top of the mouth bar.The aspect ratio ranges from 25 to 50 and there is a double logarithmic linear positive relationship between the width and thickness for the bar finger, which is controlled by base-level changing in study area. For the bar finger, injection and production in the same distributary channel should be avoided during water flooding development. In addition, middle–upper distributary channel and undrilled mouth bar are focus of tapping remaining oil.
文摘股票市场的准确预测对投资者和监管层而言都具有重要意义.在金融时间序列预测研究中分解-集成框架被广泛使用,然而以往研究中针对分解后的模态分量多数采用单一历史数据预测未来的思路,往往忽略了其他低频异质数据对分量的影响.本文融合了分解-集成与混频数据采样思想,提出EEMD-Mixed Frequency CNN-BiLSTM-Attention/LSTM-LSTM(EE-MFCBA/L-L)股指收益率预测模型,通过EEMD将股指收益率分解为若干不同频率特征的分量,采用模糊熵算法识别分量频率特征,进而结合不同频率倍差的低频数据,使用MFCBA/L模型实现对模态分量的预测,最后采用LSTM模型非线性集成各分量的预测结果.实证结果表明,所提出的模型可以更好地适应收益率变化特征,与传统模型相比,所提模型在预测非平稳和非线性收益率序列方面具有显著优势,具有最低预测误差和最高的方向性预测准确率.
