文本的语义表示是自然语言处理和机器学习领域的研究难点,针对目前文本表示中的语义缺失问题,基于LDA主题模型和Word2vec模型,提出一种新的文本语义增强方法Sem2vec(semantic to vector)模型。该模型利用LDA主题模型获得单词的主题分布...文本的语义表示是自然语言处理和机器学习领域的研究难点,针对目前文本表示中的语义缺失问题,基于LDA主题模型和Word2vec模型,提出一种新的文本语义增强方法Sem2vec(semantic to vector)模型。该模型利用LDA主题模型获得单词的主题分布,计算单词与其上下文词的主题相似度,作为主题语义信息融入到词向量中,代替one-hot向量输入至Sem2vec模型,在最大化对数似然目标函数约束下,训练Sem2vec模型的最优参数,最终输出增强的语义词向量表示,并进一步得到文本的语义增强表示。在不同数据集上的实验结果表明,相比其他经典模型,Sem2vec模型的语义词向量之间的语义相似度计算更为准确。另外,根据Sem2vec模型得到的文本语义向量,在多种文本分类算法上的分类结果,较其他经典模型可以提升0.58%~3.5%,同时也提升了时间性能。展开更多
In order to understand the migration law of respirable dust and gain reasonable design parameters for dust control on a super-long double-shearer fully mechanized working face, this paper describes research carried ou...In order to understand the migration law of respirable dust and gain reasonable design parameters for dust control on a super-long double-shearer fully mechanized working face, this paper describes research carried out using a numerical simulation package(Fluent) based on gas-solid coupling dispersed multiphase flow model and field measurement to research different technology modes, dust distribution law at different intervals where shearers work in opposite directions on the lower 9303 face, No. 2 Jining Mine,Yankuang Coal Mining Co. Results show that the concentration of dust 3–6 m away from the shearers working in the same directions was large, while the impact area of respirable dust near the shearer increased significantly to 5–6 m with the distance between two shearers working in opposite directions.The concentration of dust on a double-shearer face was considerably higher than that of a face with one shear under the combined effect of wind speed on the face and disturbed wind around the shearer, while the dust concentration near the shearer on the return side was considerably higher than that on the inlet side. The concentration of dust on a double-shearer face along the airflow declined slowly so that dust was hard to control. Simulation results confirmed the results of field measurement, which could provide reference for dust prevention design.展开更多
文摘文本的语义表示是自然语言处理和机器学习领域的研究难点,针对目前文本表示中的语义缺失问题,基于LDA主题模型和Word2vec模型,提出一种新的文本语义增强方法Sem2vec(semantic to vector)模型。该模型利用LDA主题模型获得单词的主题分布,计算单词与其上下文词的主题相似度,作为主题语义信息融入到词向量中,代替one-hot向量输入至Sem2vec模型,在最大化对数似然目标函数约束下,训练Sem2vec模型的最优参数,最终输出增强的语义词向量表示,并进一步得到文本的语义增强表示。在不同数据集上的实验结果表明,相比其他经典模型,Sem2vec模型的语义词向量之间的语义相似度计算更为准确。另外,根据Sem2vec模型得到的文本语义向量,在多种文本分类算法上的分类结果,较其他经典模型可以提升0.58%~3.5%,同时也提升了时间性能。
基金the National Natural Science Foundation of China (No.51404249)the Basic Research Program of Jiangsu Province (No.BK20140201)the Priority Academic Program Development of Jiangsu Higher Education Institutions for financial support provided during this research
文摘In order to understand the migration law of respirable dust and gain reasonable design parameters for dust control on a super-long double-shearer fully mechanized working face, this paper describes research carried out using a numerical simulation package(Fluent) based on gas-solid coupling dispersed multiphase flow model and field measurement to research different technology modes, dust distribution law at different intervals where shearers work in opposite directions on the lower 9303 face, No. 2 Jining Mine,Yankuang Coal Mining Co. Results show that the concentration of dust 3–6 m away from the shearers working in the same directions was large, while the impact area of respirable dust near the shearer increased significantly to 5–6 m with the distance between two shearers working in opposite directions.The concentration of dust on a double-shearer face was considerably higher than that of a face with one shear under the combined effect of wind speed on the face and disturbed wind around the shearer, while the dust concentration near the shearer on the return side was considerably higher than that on the inlet side. The concentration of dust on a double-shearer face along the airflow declined slowly so that dust was hard to control. Simulation results confirmed the results of field measurement, which could provide reference for dust prevention design.