Natural language is a language that has naturally emerged to enable communication between people. Technologies to mechanically analyze human utterance, to understand natural language for conversion into a form for understanding by computer and, reversely, to express it for understanding by people are all classified as natural language processing technologies.
To understand natural language, it is necessary to understand sentence structure (grammar), which is to determine parts of speech of the words that compose a sentence and the structures of each sentence, and to semantically understand the words. NLP is configured with the below key functions.
This is the most important part of a voice conversation system. Conversation processing technology based on natural language provides language analysis and understanding as well as conversation management and utterance generation functionality.
- Morpheme analysis
- Character strings entered as a result of voice recognition are separated into morphemes, which are the minimum semantic unit, by using regular grammar, dictionary information and the degree of combination. Then, restoration and recognition of the parts of speech are executed.
- Sentence structure analysis
- Structurally analyzes syntax, which is a rule to make a sentence or a phrase by combining morphemes.
- Analysis of meaning
- Analyzes the meaning of a sentence by examining syntax analysis result.
- Utterance generation
- Generates words or sentences for response from utterance dictionary based on the expression of meanings entered through speaker’s utterance analysis.