Three aspects of NLP analysis Technology NLP Analysis technology is divided into three levels: lexical analysis, syntactic analysis and semantic analysis.
1 Lexical analysis includes word segmentation, POS tagging, named entity recognition and Word sense disambiguation. Participle and part of speech to mark good understanding. The task of named entity recognition is to identify named entities, such as names, names, and organization names in sentences. Each named entity is composed of one or more words. The meaning disambiguation is to judge the true meaning of each and every word according to the context of the sentence.
2 Syntactic analysis is the key step in NLP, which is to transform the input sentence from the sequence form into a tree structure, so as to capture the collocation or modification relationship between the words in the sentence. At present, there are two kinds of mainstream syntactic analysis methods in the research field: Phrase structure syntax system and dependent structure syntax system. The system of dependency syntax has now become a hotspot in the study of syntactic analysis. The expression of dependency grammar is simple, easy to understand and annotation, it can easily express the semantic relationship between words, such as the relationship between the sentence composition can constitute agent, patient, time and so on. This semantic relationship can be conveniently applied to fish semantic analysis and information extraction. Dependencies can also be more efficient in implementing decoding algorithms. Syntactic structure analysis can help the upper level of semantic analysis, as well as some applications, such as machine translation, question and answer, text mining, information retrieval and so on.
3 The ultimate goal of semantic analysis is to understand the true semantics of sentence expression. In what form to express semantics has not been able to solve very well. Semantic role tagging is a relatively mature technique for shallow semantic analysis. Given a predicate in a sentence, the task of semantic role tagging is to mark the predicate's agent, patient, time, place and other parameters from the sentence. Semantic role tagging is generally done on the basis of syntactic analysis, and syntactic structure is critical to the performance of semantic role tagging.
