Input method, apparatus, and electronic device

• US 10,345,923 B2
• Filed: 08/14/2015
• Issued: 07/09/2019
• Est. Priority Date: 09/09/2014
• Status: Active Grant

• Current Assignee: Beijing Sogou Technology Development Co.,Ltd. (Tencent Holdings Limited)
• Original Assignee: Beijing Sogou Technology Development Co.,Ltd. (Tencent Holdings Limited)
• Inventors: Cui, Xin, Ren, Shangkun, Su, Qilong, Zhang, Yang
• Primary Examiner(s): Phantana-angkool, David
• Application Number:
  US15/505,263
• Publication Number:
  US 20180217676A1
• Time in Patent Office: 1,425 Days
• Field of Search: 715816
• CPC Class Codes: G06F 16/00 Information retrieval; Database structures therefor; File system structures therefor
• CPC Class Codes: G06F 16/3343 using phonetics
• CPC Class Codes: G06F 16/3349 Reuse of stored results of previous queries
• CPC Class Codes: G06F 3/018 Input/output arrangements for oriental characters
• CPC Class Codes: G06F 3/0236 using selection techniques to select from displayed items
• CPC Class Codes: G06F 3/0237 using prediction or retrieval techniques
• CPC Class Codes: G06F 3/04886 by partitioning the display area of the touch-screen or the surface of the digitising tablet into independently controllable areas, e.g. virtual keyboards or menus
• CPC Class Codes: G06F 40/274 Converting codes to words; Guess-ahead of partial word inputs

Abstract

The present disclosure provides an input method, an apparatus and an electronic device. The input method comprises: acquiring historical dialog information about a dialog window; extracting characteristic information in the historical dialog information, wherein the characteristic information comprises a keyword; determining prediction information, wherein the prediction information comprises the keyword; based on the prediction information, acquiring an on-screen candidate word queue in an input text box in the dialog window; and outputting the on-screen candidate word queue.

Claims

1. An input method, comprising:

acquiring, by an input method application, historical dialogue information about a dialogue window of a third-party application, the historical dialogue information being information that has been generated and recorded in the dialogue window of the third-party application;

extracting, by the input method application, characteristic information in the historical dialogue information, wherein the characteristic information comprises a keyword;

determining, by the input method application, prediction information, wherein the prediction information comprises the keyword;

based on the prediction information, acquiring, by the input method application, an on-screen candidate word queue for an input text box in the dialogue window; and

outputting, by the input method application, the on-screen candidate word queue containing word candidates to be selected and entered onto the input text box in the dialogue window of the third-party application.

2. The input method according to claim 1, wherein acquiring historical dialogue information about a dialogue window further includes:

acquiring the historical dialogue information about the dialog window within a preset history timeframe.

3. The input method according to claim 1, wherein determining prediction information further includes:

determining whether the input text box of the dialog window includes text information;

when the input text box of the dialog window does not include the text information, determining the prediction information to be the keyword;

when the input text box of the dialog window includes the text information, determining whether there is a topic shift between the text information and the keyword; and

when there is no topic shift between the text information and the keyword, determining the prediction information to be the keyword and the text information.

4. The input method according to claim 3, wherein determining whether there is a topic shift between the text information and the keyword further includes:

determining whether there is an overlap between a first set of scenario feature tags hit by the text information and a second set of scenario feature tags hit by the keyword, wherein a set of the scenario feature tags hit by a word is pre-categorized and pre-determined based on a meaning or a definition of the word;

when there is an overlap between the first set of the scenario feature tags hit by the text information and the second set of the scenario feature tags hit by the keyword, there is no topic shift between the between the text information and the keyword; and

when there is no overlap between the first set of scenario feature tags hit by the text information and the second set of scenario feature tags hit by the keyword, there is a topic shift between the between the text information and the keyword.

5. The input method according to claim 3, wherein based on the prediction information, acquiring an on-screen candidate word queue in an input text box in the dialogue window further includes:

determining the scenario feature tags hit by the prediction information;

ranking the scenario feature tags;

based on the prediction information, searching an associated candidate word bank of the scenario feature tags to obtain on-screen candidate words in; and

based on the obtained on-screen candidate words, acquiring the on-screen candidate word queue in the input text box.

6. The input method according to claim 5, wherein ranking the scenario feature tags further includes:

ranking the scenario feature tags, according to a number of the pieces of prediction information hitting each scene feature tag, and time when the prediction information appears in the dialogue window.

7. The input method according to claim 5, wherein when the characteristic information extracted includes an identification word, before based on the obtained on-screen candidate words, acquiring the on-screen candidate word queue in the input text box, further includes:

categorizing and saving the identification word according to the scenarios feature tags; and

based on a rank of the scene feature tags, obtaining the identification word with the corresponding scene feature tag, wherein the obtained identification word is taken as the on-screen candidate word in the input text box.

8. The input method according to claim 5, wherein before based on the obtained on-screen candidate words, acquiring the on-screen candidate word queue in the input text box, further includes:

based on the prediction information, searching for the on-screen candidate words in an associated candidate word bank of the language model,

wherein when the prediction information includes the keyword, the keyword is taken as a left-gram, and the language model is a bigram language model,

when the prediction information includes the keyword and the text information, the keyword is taken as the left-gram, the text information is taken as a middle-gram, and the language model is a trigram language model, or the text information is taken as the left-gram, and the language model is the bigram language model.

9. An input apparatus, comprising at least one processor, the at least one processor being configured for:

acquiring, by an input method application, historical dialogue information about a dialogue window of a third-party application, the historical dialogue information being information that has been generated and recorded in the dialogue window of the third-party application;

extracting, by the input method application, characteristic information in the historical dialogue information, wherein the characteristic information comprises a keyword;

determining, by the input method application, prediction information, wherein the prediction information comprises the keyword;

based on the prediction information, acquiring, by the input method application, an on-screen candidate word queue for an input text box in the dialogue window; and

outputting, by the input method application, the on-screen candidate word queue containing word candidates to be selected and entered onto the input text box in the dialogue window of the third-party application.

10. The input apparatus according to claim 9, wherein the at least one processor is further configured for:

acquiring the historical dialogue information about the dialog window within a preset history timeframe;

or

acquiring a preset number of pieces of the historical dialogue information about the dialog window.

11. The input apparatus according to claim 9, wherein the at least one processor is further configured for:

determining whether an input text box of the dialog window includes any text information;

when the input text box of the dialog window does not include the text information, determining the prediction information to be the keyword;

when the input text box of the dialog window includes the text information, determining whether there is a topic shift between the text information and the keyword; and

when there is no topic shift between the text information and the keyword, determining the prediction information to be the keyword and the text information.

12. The input apparatus according to claim 11, wherein the at least one processor is further configured for:

determining whether there is an overlap between a first set of scenario feature tags hit by the text information and a second set of scenario feature tags hit by the keyword, wherein a set of the scenario feature tags hit by a word is pre-categorized and pre-determined based on a meaning or a definition of the word;

when there is an overlap between the first set of the scenario feature tags hit by the text information and the second set of the scenario feature tags hit by the keyword, there is no topic shift between the between the text information and the keyword; and

when there is no overlap between the first set of scenario feature tags hit by the text information and the second set of scenario feature tags hit by the keyword, there is a topic shift between the between the text information and the keyword.

13. The input apparatus according to claim 11, wherein the at least one processor is further configured for:

determining scenario feature tags hit by the prediction information;

ranking the scenario feature tags;

searching an associated candidate word bank of the scenario feature tags to obtain on-screen candidate words in; and

based on the obtained on-screen candidate words, acquiring an on-screen candidate word queue in the input text box.

14. The input apparatus according to claim 13, wherein the at least one processor is further configured for:

ranking the scenario feature tags, according to a number of the pieces of prediction information hitting each scene feature tag, and time when the prediction information appears in the dialogue window.

15. The input apparatus according to claim 13, wherein the at least one processor is further configured for:

before acquiring the on-screen candidate word queue in the input text box, categorizing and saving the identification words according to the scenarios feature tags; and

based on a rank of the scene feature tags, obtaining the identification word with the corresponding scene feature tag, and taking the obtained identification word as the on-screen candidate word in the input text box.

16. The input apparatus according to claim 13, the at least one processor is further configured for:

before acquiring the on-screen candidate word queue in the input text box, based on the prediction information, searching for the on-screen candidate words in an associated candidate word bank of the language model,

wherein when the prediction information includes the keyword, the keyword is taken as a left-gram, and the language model is a bigram language model,

when the prediction information includes the keyword and the text information, the keyword is taken as the left-gram, the text information is taken as a middle-gram, and the language model is a trigram language model, or the text information is taken as the left-gram, and the language model is the bigram language model.

17. A non-transitory computer-readable medium comprising a computer readable code for, when being executed by at least one processor of a mobile terminal, performing an input method, wherein the method comprising:

acquiring, by an input method application, historical dialogue information about a dialogue window of a third-party application, the historical dialogue information being information that has been generated and recorded in the dialogue window of the third-party application;

extracting, by the input method application, characteristic information in the historical dialogue information, wherein the characteristic information comprises a keyword;

determining, by the input method application, prediction information, wherein the prediction information comprises the keyword;

based on the prediction information, acquiring, by the input method application, an on-screen candidate word queue for an input text box in the dialogue window; and

outputting, by the input method application, the on-screen candidate word queue containing word candidates to be selected and entered onto the input text box in the dialogue window of the third-party application.

18. The non-transitory computer-readable medium according to claim 17, wherein acquiring historical dialogue information about a dialogue window further includes:

acquiring the historical dialogue information about the dialog window within a preset history timeframe;

or

acquiring a preset number of pieces of the historical dialogue information about the dialog window.

19. The non-transitory computer-readable medium according to claim 17, wherein determining prediction information further includes:

determining whether the input text box of the dialog window includes text information;

when the input text box of the dialog window does not include the text information, determining the prediction information to be the keyword;

when the input text box of the dialog window includes the text information, determining whether there is a topic shift between the text information and the keyword; and

when there is no topic shift between the text information and the keyword, determining the prediction information to be the keyword and the text information.