Method and apparatus for recommending word

Patent metadata

Field	Value
Publication number	US-10846477-B2
Application number	US-201815954931-A
Country	US
Kind code	B2
Filing date	Apr 17, 2018
Priority date	May 16, 2017
Publication date	Nov 24, 2020
Grant date	Nov 24, 2020

Abstract

Provided is a device including a memory storing information about sequences of a plurality of registered words; an input unit comprising input circuitry configured to receive an input of a text comprising a first eojeol not belonging to the plurality of registered words, wherein, in the first eojeol, a first word is attached to a first registered word that belongs to the plurality of registered words; and a controller configured to detect the first registered word from the first eojeol, to determine a predicted eojeol to be input after the text, based on the information about the sequences of the plurality of registered words and the detected first registered word and to control a display to display the predicted eojeol.

First claim

What is claimed is:

1. A device comprising:
a display;
a memory storing information about sequences about a plurality of registered words; and
at least one processor configured to:
receive a user input including a text comprising a first combined word,
detect a first portion corresponding to at least one word of the plurality of registered words, from the first combined word,
determine a plurality of predicted words to be input in succession to the text via a neural network, based on the information about the sequences about the plurality of registered words and the detected first portion, including modify and refine parameters for the neural network so as to train the neural network so that next words can be output as an output value from the neural network,
control the display to display a menu for receiving an input selecting one of the plurality of the predicted words predicted via the neural network
obtain an input selecting a word from among the plurality of predicted words, and
input the selected word in succession to the text via the neural network,
wherein the first combined word comprises the first portion corresponding to at least one word of the plurality of registered words and a second portion not corresponding to any words of the plurality of registered words.

2. The device of claim 1 , wherein the at least one processor is further configured to detect the first portion corresponding to at least one word of the plurality of registered words from the first combined word by applying a plurality of filters corresponding to the plurality of registered words to a plurality of syllables forming the first combined word.

3. The device of claim 1 , wherein the at least one processor is further configured to detect at least one registered word comprising the first portion from the text, and to determine a registered word as the predicted word, based on the information about the sequences about the plurality of registered words, wherein the registered word has a highest probability of being input after a sequence about the at least one registered word from among the plurality of registered words.

4. The device of claim 3 , wherein the at least one processor is further configured to determine the predicted word using a neural network algorithm, wherein the predicted word has a highest probability of being input after the sequence about the at least one registered word from among the plurality of registered words.

5. The device of claim 4 , wherein the at least one processor is further configured to:
receive an input of inputting a next word of the text, and
modify and refine parameters of the neural network algorithm for the neural network so that, when the sequence about the at least one registered word detected from the text is input as an input value to the neural network, the next word is output as an output value from the neural network.

6. The device of claim 1 , wherein
the at least one processor is further configured to determine a plurality of predicted words to be input in succession to the text, based on the information about the sequences about the plurality of registered words and the detected first portion, and
the display is further configured to display a menu for receiving an input selecting one of the plurality of predicted words, and when an input selecting an word from among the plurality of predicted words is received, the at least one processor is further configured to input the selected word in succession to the text.

7. The device of claim 1 , wherein the at least one processor is further configured to store the first portion as an affix with respect to the at least one word of the plurality of registered words, and when the at least one word of the plurality of registered words is determined as the predicted word, the at least one processor is further configured to control the display to display a menu for receiving an input selecting one of the at least one word of the plurality of registered words and the first combined word.

8. A method of recommending a word, the method comprising:
storing, in memory, information about sequences about a plurality of registered words;
receiving, by at least one processor, a user input including a text comprising a first combined word;
detecting a first portion corresponding to at least one word of the plurality of registered words via a neural network, from the first combined word, and determining a predicted word to be input in succession to the text, based on the information about the sequences about the plurality of registered words and the detected first portion, including modifying and refining parameters for the neural network so as to train the neural network so that next words can be output as an output value from the neural network; and
displaying, on a display, the predicted word,
wherein the first combined word comprises the first portion corresponding to at least one word of the plurality of registered words and a second portion not corresponding to any words of the plurality of registered words.

9. The method of claim 8 , wherein the detecting of the first portion corresponding to at least one word of the plurality of registered words comprises detecting the first portion corresponding to at least one word of the plurality of registered words from the first combined word by applying a plurality of filters corresponding to the plurality of registered words to a plurality of syllables forming the first combined word.

10. The method of claim 8 , wherein the determining of the predicted word comprises detecting at least one registered word comprising the first portion from the text, and determining a registered word as the predicted word, based on the information about the sequences about the plurality of registered words, wherein the registered word has a highest probability of being input after a sequence about the at least one registered word from among the plurality of registered words.

11. The method of claim 10 , wherein the determining of the predicted word comprises determining the predicted word using a neural network algorithm, wherein the predicted word has a highest probability of being input after the sequence about the at least one registered word from among the plurality of registered words.

12. The method of claim 11 , further comprising:
receiving, by the at least one processor, an input including a next word of the text; and
modifying and refining parameters of the neural network algorithm for the neural network so that, when the sequence about the at least one registered word detected from the text is input as an input value to the neural network, the next word is output as an output value from the neural network.

13. The method of claim 8 , wherein the determining of the predicted word comprises determining the predicted word to be input in succession to the text, based on the first portion detected from the first combined word and at least one morpheme included in the first combined word.

14. The method of claim 8 , wherein the determining of the predicted word comprises determining a plurality of predicted words to be input after the text, based on the information about the sequences about the plurality of registered words and the detected first portion, and
the method further comprises displaying, on the display, a menu for receiving an input selecting one of the plurality of predicted words, and when an input selecting an word from among the plurality of predicted words is received, inputting the selected word in succession to the text.

15. The method of claim 8 , further comprising:
storing, in the m

Assignees

• Samsung Electronics Co Ltd

Inventors

• Lee Hae-Jun
• Yu Seung-Hak

Classifications

• G06F3/0482

  Interaction with lists of selectable items, e.g. menus · CPC title

• G06F3/0237 Primary

  using prediction or retrieval techniques · CPC title

• G06F40/53

  Processing of non-Latin text (kana-to-kanji conversion G06F40/129; vowelisation G06F40/232) · CPC title

• G06F40/268

  Morphological analysis · CPC title

• G06F40/253

  Grammatical analysis; Style critique · CPC title