It mainly focuses on two levels of the language: daily conversation and semantic parameters.

This paper presents a method of using soft pattern to discover semantic classes.

Image semantic classification based on multi-features;

This paper analyzes and studies their semantic classification and feature,and pragmatic analysis.

An image annotation method based on the semantic classification is put forward.

Through metaphoricalization,the temporal words which serve as the temporal background originally change into the entities and participants of the events,and thus serve as subjects of different semantic types.

The characteristics lie in the aspects of semantic types, morpheme types and whether the compounds are used as independent sentences.

From their syntactic formation, there are some clear differences,not only between the composed componems Of whole construction, but also between the degrees of complication,the linear length and the semantic type in items of comparison,and the semantic function in items of conclusion.

Web Information Retrieval System Based on Classification Semantics;

In order to improve the performance of search engine this paper applies a new technology of web-page classification to the existing search engine and designs a search engine system which is based on classification semantics——SECS system.