Motivated by a systematic analysis of Chinese semantic relationships, we constructed a Chinese semantic framework based on surface syntactic relationships, deep semantic relationships and feature structure to express dependencies between lexical meanings and conceptual structures, and relations that underlie those lexical meanings. Analyzing the semantic representations of 10000 Chinese sentenc...

Semantic relations of Chinese verb-complement structure are complicated, which is difficult to analyze semantic relations in NLP. This paper proposes a novel model based on “the Feature Structure theory” and applies it in representing the semantic relations among the four components, which are subject, verb, object and complement. We annotated the fifteen types of semantic relations, and compar...

Semantic feature norms, originally utilized in the field of psycholinguistics as a tool for studying human semantic representation and computation, have recently attracted the attention of some NLP/IR researchers who wish to improve their task performances. However, currently available semantic feature norms are, by nature, not well-structured, making them difficult to integrate into existing r...

How verbal and nonverbal visuoperceptual input connects to semantic knowledge is a core question in visual and cognitive neuroscience, with significant clinical ramifications. In an event-related functional magnetic resonance imaging (fMRI) experiment we determined how cosine similarity between fMRI response patterns to concrete words and pictures reflects semantic clustering and semantic dista...

Text clustering and classification are two main tasks of text mining. Feature selection plays the key role in the quality of the clustering and classification results. Although word-based features such as term frequency-inverse document frequency (TF-IDF) vectors have been widely used in different applications, their shortcoming in capturing semantic concepts of text motivated researches to use...

Twitter has brought much attention recently as a hot research topic in the domain of sentiment analysis. Training sentiment classifiers from tweets data often faces the data sparsity problem partly due to the large variety of short and irregular forms introduced to tweets because of the 140-character limit. In this work we propose using two different sets of features to alleviate the data spars...