Two lines of research on eye movements in reading are summarized. One line of research examines how adult readers identify compound words during reading. The other line of research deals with how a specific reading go...Two lines of research on eye movements in reading are summarized. One line of research examines how adult readers identify compound words during reading. The other line of research deals with how a specific reading goal influences the way long expository texts are read. Both lines of research are conducted using Finnish as the source language. With respect to the first research question, it is demonstrated that compound words are recognized either holistically or via their components, depending on the length of the compound word. Readers begin to process whatever information is readily available in the foveal vision(i.e., either the whole-word form or the initial component). The second line of research demonstrates that(1)a specific reading goal is capable of exerting an early effect on readers’ eye fixation patterns,(2)time course analyses based on eye movement patterns can reveal interesting individual differences, and(3)working memory capacity is linked to the efficiency to strategically allocate attention as well as to encode information to and retrieve it from the long-term memory. It is concluded that the eye-tracking technique is an excellent research tool to tap into the workings of the human mind during the comprehension of written texts.展开更多
Text Rank is a popular tool for obtaining words or phrases that are important for many Natural Language Processing (NLP) tasks. This paper presents a practical approach for Text Rank domain specific using Field Associ...Text Rank is a popular tool for obtaining words or phrases that are important for many Natural Language Processing (NLP) tasks. This paper presents a practical approach for Text Rank domain specific using Field Association (FA) words. We present the keyphrase separation technique not for a single document, although for a particular domain. The former builds a specific domain field. The second collects a list of ideal FA terms and compounds FA terms from the specific domain that are considered to be contender keyword phrases. Therefore, we combine two-word node weights and field tree relationships into a new approach to generate keyphrases from a particular domain. Studies using the changed approach to extract key phrases demonstrate that the latest techniques including FA terms are stronger than the others that use normal words and its precise words reach 90%.展开更多
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
为了对饮食文本信息高效分类,建立一种基于word2vec和长短期记忆网络(Long-short term memory,LSTM)的分类模型。针对食物百科和饮食健康文本特点,首先利用word2vec实现包含语义信息的词向量表示,并解决了传统方法导致数据表示稀疏及维...为了对饮食文本信息高效分类,建立一种基于word2vec和长短期记忆网络(Long-short term memory,LSTM)的分类模型。针对食物百科和饮食健康文本特点,首先利用word2vec实现包含语义信息的词向量表示,并解决了传统方法导致数据表示稀疏及维度灾难问题,基于K-means++根据语义关系聚类以提高训练数据质量。由word2vec构建文本向量作为LSTM的初始输入,训练LSTM分类模型,自动提取特征,进行饮食宜、忌的文本分类。实验采用48 000个文档进行测试,结果显示,分类准确率为98.08%,高于利用tf-idf、bag-of-words等文本数值化表示方法以及基于支持向量机(Support vector machine,SVM)和卷积神经网络(Convolutional neural network,CNN)分类算法结果。实验结果表明,利用该方法能够高质量地对饮食文本自动分类,帮助人们有效地利用健康饮食信息。展开更多
文摘Two lines of research on eye movements in reading are summarized. One line of research examines how adult readers identify compound words during reading. The other line of research deals with how a specific reading goal influences the way long expository texts are read. Both lines of research are conducted using Finnish as the source language. With respect to the first research question, it is demonstrated that compound words are recognized either holistically or via their components, depending on the length of the compound word. Readers begin to process whatever information is readily available in the foveal vision(i.e., either the whole-word form or the initial component). The second line of research demonstrates that(1)a specific reading goal is capable of exerting an early effect on readers’ eye fixation patterns,(2)time course analyses based on eye movement patterns can reveal interesting individual differences, and(3)working memory capacity is linked to the efficiency to strategically allocate attention as well as to encode information to and retrieve it from the long-term memory. It is concluded that the eye-tracking technique is an excellent research tool to tap into the workings of the human mind during the comprehension of written texts.
文摘Text Rank is a popular tool for obtaining words or phrases that are important for many Natural Language Processing (NLP) tasks. This paper presents a practical approach for Text Rank domain specific using Field Association (FA) words. We present the keyphrase separation technique not for a single document, although for a particular domain. The former builds a specific domain field. The second collects a list of ideal FA terms and compounds FA terms from the specific domain that are considered to be contender keyword phrases. Therefore, we combine two-word node weights and field tree relationships into a new approach to generate keyphrases from a particular domain. Studies using the changed approach to extract key phrases demonstrate that the latest techniques including FA terms are stronger than the others that use normal words and its precise words reach 90%.
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。
