Electronic device and natural language analysis method thereof

摘要

A language analysis method for an electronic device storing a basic corpus and a temporary corpus is provided. The language analysis method includes steps of receiving user inputs and generating signals; converting signals into textualized information; analyzing the textualized information; obtaining a first understanding result according to the basic corpus, the vocabulary segmentation results, and the sentence segmentation results; determining whether the first understanding result is an appropriate understanding according to the context; determining one or more anaphoric vocabularies when the first understanding result is an inappropriate understanding; determining a temporary understanding result of the one or more anaphoric vocabularies and a second understanding result of the textualized information according to the context; and determining a reply for the textualized information, according to the second understanding result, the basic corpus, and the temporary corpus. An electronic device using the language analysis method is also provided.

主权项

1. A natural language analysis method for an electronic device storing a basic corpus and a temporary corpus, the basic corpus recording a vast amount of vocabularies, words and phrases and a frequency of use of each word and each phrase, the temporary corpus recording the mapping relationship between at least one substitute vocabulary and at least one temporary meaning, the method comprising:
generating signals in response to a user's input; converting the signals into a textualized message in a predetermined language; segmenting the textualized message into at least one vocabulary, and obtaining at least one vocabularized segment comprising the at least one vocabulary; segmenting the textualized message based on the vocabularized segments and a sentence construction rule, and obtaining at least one sentence segment; retrieving the frequency of use of each segmented vocabulary from the basic corpus, obtaining a first language understanding result of the textualized message by analyzing the at least one vocabularized segments and the at least one sentence segments, based on the retrieved frequency of use of each segmented vocabulary; selecting a plurality of textualized messages consecutively converted within a predetermined time period, the selected textualized messages including said textualized message which is segmented in the segmenting step; analyzing of the selected textualized messages using a contextual understanding method, so as to determine whether the first language understanding result is a reasonable understanding; if the first language understanding result is an unreasonable understanding, determining at least one substitute vocabulary causing the unreasonable understanding, determining whether the temporary corpus records the determined at least one substitute vocabulary; if the temporary corpus does not have a record of one or more of the determined at least one substitute vocabulary, determining temporary meaning of each determined substitute vocabulary by analyzing the context comprising the determined substitute vocabulary to obtain a second understanding result of the textualized message, and updating the temporary corpus by storing the mapping relationship between the determined at least one substitute vocabulary and the temporary meaning therein; and determining a reply message for the textualized message based on a reasonable language understanding result, the temporary corpus, and the basic corpus.