近年来,随着计算能力的不断提高和网站上大量数据的免费获取,深度学习(deep learning,DL)技术得到广泛应用,对自然语言处理领域产生了巨大的影响,其中BERT(bidirectional encoder representations from transformers,BERT)已在许多自然...近年来,随着计算能力的不断提高和网站上大量数据的免费获取,深度学习(deep learning,DL)技术得到广泛应用,对自然语言处理领域产生了巨大的影响,其中BERT(bidirectional encoder representations from transformers,BERT)已在许多自然语言处理(natural language process,NLP)任务以及其他领域中得到广泛使用。如果将BERT用于基于方面级情感分析(aspect-based sentiment analysis,ABSA)任务,通过检查用户在产品评论中表达的情感类型和情感目标来研究消费者对市场产品的看法,将大大提高产品在未来市场的地位。针对这个问题,提出了并行聚合和分层聚合2个模块,应用于ABSA的2个主要任务:方面级提取(aspect extraction,AE)和方面级情感分类(aspect sentiment classification,ASC)。这些模块利用BERT语言模型的隐藏层来产生输入序列的更深层语义表示,通过并行方式进行聚合并且进行了分类,对选定的每个隐藏层进行预测并计算损失,然后将这些损失求和以产生模型的最终损失。此外,通过使用条件随机字段(conditional random fields,CRF)解决方面级提取问题。经过研究表明,在BERT微调中应用提出的模型,可以提高BERT模型的性能。展开更多
我国正逐步制定和完善能源系统用户侧的降碳政策,住宅电采暖系统运行面临新的发展契机与挑战。为提升电采暖系统运行中电热能源利用率,实现采暖用能环节经济低碳化运行,以碳税定价政策作为环境成本价格背景,提出考虑家用电器电热特性的...我国正逐步制定和完善能源系统用户侧的降碳政策,住宅电采暖系统运行面临新的发展契机与挑战。为提升电采暖系统运行中电热能源利用率,实现采暖用能环节经济低碳化运行,以碳税定价政策作为环境成本价格背景,提出考虑家用电器电热特性的分散式电采暖集群经济低碳调控策略。首先,以电器设备运行中的电热特性作为分类依据,对室内电器运行情况进行分类聚合预测,并计算电器运行热增益作为采暖热源的补充。其次,根据分散式电采暖用户建筑和采暖设备热力学特性,构建电采暖“经济-低碳”运行优化模型,采用显式模型预测控制(explicit model predictive control,EMPC)技术对模型进行求解。最后,通过算例仿真对比分析可知,所提调控策略可用于分散式采暖用户集群的实时调控,实现电采暖系统运行经济性、低碳化目标。展开更多
In order to analyze the heterogeneity in vehicular traffic speed, a new method that integrates cluster analysis and probability distribution function fitting is presented. First, for identifying the optimal number of ...In order to analyze the heterogeneity in vehicular traffic speed, a new method that integrates cluster analysis and probability distribution function fitting is presented. First, for identifying the optimal number of clusters, the two-step cluster method is applied to analyze actual speed data, which suggests that dividing speed data into two clusters can best reflect the intrinsic patterns of traffic flows. Such information is then taken as guidance in probability distribution function fitting. The normal, skew-normal and skew-t distribution functions are used to fit the probability distribution of each cluster respectively, which suggests that the skew-t distribution has the highest fitting accuracy; the second is skew-normal distribution; the worst is normal distribution. Model analysis results demonstrate that the proposed mixture model has a better fitting and generalization capability than the conventional single model. In addition, the new method is more flexible in terms of data fitting and can provide a more accurate model of speed distribution.展开更多
In k-means clustering, we are given a set of n data points in d-dimensional space R^d and an integer k and the problem is to determine a set of k points in R^d, called centers, so as to minimize the mean squared dista...In k-means clustering, we are given a set of n data points in d-dimensional space R^d and an integer k and the problem is to determine a set of k points in R^d, called centers, so as to minimize the mean squared distance from each data point to its nearest center. In this paper, we present a simple and efficient clustering algorithm based on the k-means algorithm, which we call enhanced k-means algorithm. This algorithm is easy to implement, requiring a simple data structure to keep some information in each iteration to be used in the next iteration. Our experimental results demonstrated that our scheme can improve the computational speed of the k-means algorithm by the magnitude in the total number of distance calculations and the overall time of computation.展开更多
Aiming at the problems of low reliability and complex operation of traditional coin-tap test of composite material,this paper introduces the grey system theory and achieves better performance.The response signals of c...Aiming at the problems of low reliability and complex operation of traditional coin-tap test of composite material,this paper introduces the grey system theory and achieves better performance.The response signals of coin-tap are classified through the grey clustering based on relation analysis,and corresponding improvements are made to the calculation method of the relation degree of nearness.First,the time history of acceleration is taken as the system behavior sequence.The improved correlation calculation method is used to solve the relation degree of nearness between the sequences,and the matrix of degree of grey relation is constructed based on this.Then,the sequence groups are summarized through the matrix,and the response signals of coin-tap are qualitatively classified according to the location of the reference sequence.Finally,the defect detection of composite materials is completed without pre-testing.The test results show that the accuracy of the coin-tap test based on improved grey clustering reaches 100%,which simplifies the operation steps while ensuring the reliability of the coin-tap test results.展开更多
The food industry is evolving more towards new forms of organization much more complex and characterized by a greater degree of coordination, whether in the form of vertical integration of explicit or implicit contrac...The food industry is evolving more towards new forms of organization much more complex and characterized by a greater degree of coordination, whether in the form of vertical integration of explicit or implicit contracts between players of different levels of the industry. Therefore, the aim of this work is the search for mechanisms that can provide value to the production phase to better increase competitiveness of the sector. For the first time, in fact, discussion about food chains have as reference a recognized legal entity, which is the integrated projects of food chain as a result of actions of agricultural policy at community, national and regional levels. The methodology is related to two steps： the administration of questionnaires to the three companies participating in food chain partnerships that have proposed a draft of integrated design of food chain in response to the notice of the Apulia region for the submission of the integrated projects of the food chain; and a cluster analysis in the wine sector of the Italian regions. The results showed, thanks to Network Analysis, the importance for the chain development of relationships formed by market relations and cooperation relations （formal and informal） and the need for more actions for the enhancement of products by research and development activities.展开更多
文摘近年来,随着计算能力的不断提高和网站上大量数据的免费获取,深度学习(deep learning,DL)技术得到广泛应用,对自然语言处理领域产生了巨大的影响,其中BERT(bidirectional encoder representations from transformers,BERT)已在许多自然语言处理(natural language process,NLP)任务以及其他领域中得到广泛使用。如果将BERT用于基于方面级情感分析(aspect-based sentiment analysis,ABSA)任务,通过检查用户在产品评论中表达的情感类型和情感目标来研究消费者对市场产品的看法,将大大提高产品在未来市场的地位。针对这个问题,提出了并行聚合和分层聚合2个模块,应用于ABSA的2个主要任务:方面级提取(aspect extraction,AE)和方面级情感分类(aspect sentiment classification,ASC)。这些模块利用BERT语言模型的隐藏层来产生输入序列的更深层语义表示,通过并行方式进行聚合并且进行了分类,对选定的每个隐藏层进行预测并计算损失,然后将这些损失求和以产生模型的最终损失。此外,通过使用条件随机字段(conditional random fields,CRF)解决方面级提取问题。经过研究表明,在BERT微调中应用提出的模型,可以提高BERT模型的性能。
文摘我国正逐步制定和完善能源系统用户侧的降碳政策,住宅电采暖系统运行面临新的发展契机与挑战。为提升电采暖系统运行中电热能源利用率,实现采暖用能环节经济低碳化运行,以碳税定价政策作为环境成本价格背景,提出考虑家用电器电热特性的分散式电采暖集群经济低碳调控策略。首先,以电器设备运行中的电热特性作为分类依据,对室内电器运行情况进行分类聚合预测,并计算电器运行热增益作为采暖热源的补充。其次,根据分散式电采暖用户建筑和采暖设备热力学特性,构建电采暖“经济-低碳”运行优化模型,采用显式模型预测控制(explicit model predictive control,EMPC)技术对模型进行求解。最后,通过算例仿真对比分析可知,所提调控策略可用于分散式采暖用户集群的实时调控,实现电采暖系统运行经济性、低碳化目标。
基金The National Science Foundation by Changjiang Scholarship of Ministry of Education of China(No.BCS-0527508)the Joint Research Fund for Overseas Natural Science of China(No.51250110075)+1 种基金the Natural Science Foundation of Jiangsu Province(No.BK200910046)the Postdoctoral Science Foundation of Jiangsu Province(No.0901005C)
文摘In order to analyze the heterogeneity in vehicular traffic speed, a new method that integrates cluster analysis and probability distribution function fitting is presented. First, for identifying the optimal number of clusters, the two-step cluster method is applied to analyze actual speed data, which suggests that dividing speed data into two clusters can best reflect the intrinsic patterns of traffic flows. Such information is then taken as guidance in probability distribution function fitting. The normal, skew-normal and skew-t distribution functions are used to fit the probability distribution of each cluster respectively, which suggests that the skew-t distribution has the highest fitting accuracy; the second is skew-normal distribution; the worst is normal distribution. Model analysis results demonstrate that the proposed mixture model has a better fitting and generalization capability than the conventional single model. In addition, the new method is more flexible in terms of data fitting and can provide a more accurate model of speed distribution.
文摘In k-means clustering, we are given a set of n data points in d-dimensional space R^d and an integer k and the problem is to determine a set of k points in R^d, called centers, so as to minimize the mean squared distance from each data point to its nearest center. In this paper, we present a simple and efficient clustering algorithm based on the k-means algorithm, which we call enhanced k-means algorithm. This algorithm is easy to implement, requiring a simple data structure to keep some information in each iteration to be used in the next iteration. Our experimental results demonstrated that our scheme can improve the computational speed of the k-means algorithm by the magnitude in the total number of distance calculations and the overall time of computation.
基金National Key Research and Development Project of China(No.2018YFB1701200)。
文摘Aiming at the problems of low reliability and complex operation of traditional coin-tap test of composite material,this paper introduces the grey system theory and achieves better performance.The response signals of coin-tap are classified through the grey clustering based on relation analysis,and corresponding improvements are made to the calculation method of the relation degree of nearness.First,the time history of acceleration is taken as the system behavior sequence.The improved correlation calculation method is used to solve the relation degree of nearness between the sequences,and the matrix of degree of grey relation is constructed based on this.Then,the sequence groups are summarized through the matrix,and the response signals of coin-tap are qualitatively classified according to the location of the reference sequence.Finally,the defect detection of composite materials is completed without pre-testing.The test results show that the accuracy of the coin-tap test based on improved grey clustering reaches 100%,which simplifies the operation steps while ensuring the reliability of the coin-tap test results.
文摘The food industry is evolving more towards new forms of organization much more complex and characterized by a greater degree of coordination, whether in the form of vertical integration of explicit or implicit contracts between players of different levels of the industry. Therefore, the aim of this work is the search for mechanisms that can provide value to the production phase to better increase competitiveness of the sector. For the first time, in fact, discussion about food chains have as reference a recognized legal entity, which is the integrated projects of food chain as a result of actions of agricultural policy at community, national and regional levels. The methodology is related to two steps： the administration of questionnaires to the three companies participating in food chain partnerships that have proposed a draft of integrated design of food chain in response to the notice of the Apulia region for the submission of the integrated projects of the food chain; and a cluster analysis in the wine sector of the Italian regions. The results showed, thanks to Network Analysis, the importance for the chain development of relationships formed by market relations and cooperation relations （formal and informal） and the need for more actions for the enhancement of products by research and development activities.
