精准的电力负荷预测有利于保障电力系统的安全、经济运行。针对现行预测算法存在的预测准确度低、模型耗时长等问题,提出一种基于随机森林(random forest,RF)算法和粗糙集理论(rough set theory,RST)的改进型深度学习(deeplearning, DL...精准的电力负荷预测有利于保障电力系统的安全、经济运行。针对现行预测算法存在的预测准确度低、模型耗时长等问题,提出一种基于随机森林(random forest,RF)算法和粗糙集理论(rough set theory,RST)的改进型深度学习(deeplearning, DL)短期负荷预测模型(RF-DL-RST)。该模型首先基于历史数据,利用随机森林算法提取影响负荷预测的关键特征量;然后将关键特征量和历史负荷值作为深度神经网络的输入、输出项进行训练,并通过粗糙集理论修正预测结果。最后,通过算例进行仿真验证,结果表明,该模型的预测准确度比单一的深度学习模型及不进行预测修正的模型更高。展开更多
高效准确的短期电力负荷预测对提升新型电力系统经济运行十分重要。针对极端天气场景下负荷预测数据量较少、随机性较强的特点,提出一种基于张量低秩补全算法的短期负荷预测模型,并选取极端高温场景展开研究。首先,给出极端天气定义,并...高效准确的短期电力负荷预测对提升新型电力系统经济运行十分重要。针对极端天气场景下负荷预测数据量较少、随机性较强的特点,提出一种基于张量低秩补全算法的短期负荷预测模型,并选取极端高温场景展开研究。首先,给出极端天气定义,并基于改进型炎热指数和气温两项指标进行数据筛选;其次,提出一种基于张量的负荷数据补全模型,补全缺失数据;然后,通过Pearson相关性分析筛选输入特征量,构建基于长短时记忆(long short term memory, LSTM)网络和粗糙集理论(rough set theory, RST)的LSTM-RST短期负荷预测模型;最后,以苏州某地实际负荷数据设置算例进行验证,仿真结果表明,所提短期负荷预测方法具有较高的准确性。展开更多
This paper proposes a clustering technique that minimizes the need for subjective human intervention and is based on elements of rough set theory （RST）. The proposed algorithm is unified in its approach to clusterin...This paper proposes a clustering technique that minimizes the need for subjective human intervention and is based on elements of rough set theory （RST）. The proposed algorithm is unified in its approach to clustering and makes use of both local and global data properties to obtain clustering solutions. It handles single-type and mixed attribute data sets with ease. The results from three data sets of single and mixed attribute types are used to illustrate the technique and establish its efficiency.展开更多
The paper discusses non-probabilistic approaches for uncertainty treatment in structure reliability analysis. Based on rough set theory, the uncertain parameters of structures are expressed by rough variables, the str...The paper discusses non-probabilistic approaches for uncertainty treatment in structure reliability analysis. Based on rough set theory, the uncertain parameters of structures are expressed by rough variables, the structure reliability index is computed by rough function and metric. This new methodology for structural reliability is proved to be valid and efficient using theory analysis and examples of practical application.展开更多
A new intelligent method for disease diagnosis based on rough set theory （RST） and the relevance vector machine （RVM） for classification is presented as the rough relevance vector machine （RRVM）. The RRVM mixes ...A new intelligent method for disease diagnosis based on rough set theory （RST） and the relevance vector machine （RVM） for classification is presented as the rough relevance vector machine （RRVM）. The RRVM mixes rough set＇s strong rule extraction ability with the excellent classification ability of the relevance vector machine through preprocessing initial information, reducing data, and training the relevance vector machine. Compared with traditional intelligence methods such as neural network（NN）, support vector machine（SVM）, and relevance vector machine （RVM）, this method manages to identify disease samples objectively and effectively with less transcendental information.展开更多
文摘精准的电力负荷预测有利于保障电力系统的安全、经济运行。针对现行预测算法存在的预测准确度低、模型耗时长等问题,提出一种基于随机森林(random forest,RF)算法和粗糙集理论(rough set theory,RST)的改进型深度学习(deeplearning, DL)短期负荷预测模型(RF-DL-RST)。该模型首先基于历史数据,利用随机森林算法提取影响负荷预测的关键特征量;然后将关键特征量和历史负荷值作为深度神经网络的输入、输出项进行训练,并通过粗糙集理论修正预测结果。最后,通过算例进行仿真验证,结果表明,该模型的预测准确度比单一的深度学习模型及不进行预测修正的模型更高。
文摘高效准确的短期电力负荷预测对提升新型电力系统经济运行十分重要。针对极端天气场景下负荷预测数据量较少、随机性较强的特点,提出一种基于张量低秩补全算法的短期负荷预测模型,并选取极端高温场景展开研究。首先,给出极端天气定义,并基于改进型炎热指数和气温两项指标进行数据筛选;其次,提出一种基于张量的负荷数据补全模型,补全缺失数据;然后,通过Pearson相关性分析筛选输入特征量,构建基于长短时记忆(long short term memory, LSTM)网络和粗糙集理论(rough set theory, RST)的LSTM-RST短期负荷预测模型;最后,以苏州某地实际负荷数据设置算例进行验证,仿真结果表明,所提短期负荷预测方法具有较高的准确性。
文摘This paper proposes a clustering technique that minimizes the need for subjective human intervention and is based on elements of rough set theory （RST）. The proposed algorithm is unified in its approach to clustering and makes use of both local and global data properties to obtain clustering solutions. It handles single-type and mixed attribute data sets with ease. The results from three data sets of single and mixed attribute types are used to illustrate the technique and establish its efficiency.
基金supported by National Natural Science Foundation of China under Grant No.61373112Special Project of Scientific Research of Education Department of Shaanxi Provincial Government No.11JK0967
文摘The paper discusses non-probabilistic approaches for uncertainty treatment in structure reliability analysis. Based on rough set theory, the uncertain parameters of structures are expressed by rough variables, the structure reliability index is computed by rough function and metric. This new methodology for structural reliability is proved to be valid and efficient using theory analysis and examples of practical application.
基金Supported by the National Natural Science Foundation of China (70771708)
文摘A new intelligent method for disease diagnosis based on rough set theory （RST） and the relevance vector machine （RVM） for classification is presented as the rough relevance vector machine （RRVM）. The RRVM mixes rough set＇s strong rule extraction ability with the excellent classification ability of the relevance vector machine through preprocessing initial information, reducing data, and training the relevance vector machine. Compared with traditional intelligence methods such as neural network（NN）, support vector machine（SVM）, and relevance vector machine （RVM）, this method manages to identify disease samples objectively and effectively with less transcendental information.
