User-level personalized text abstract generation method and system
A user-level, text-based technology, applied in unstructured text data retrieval, text database browsing/visualization, special data processing applications, etc., can solve the difficulty of user background information collection, lack of personalized text summarization methods, User-level personalization and other issues, to achieve the effect of simple and accurate acquisition methods
Active Publication Date: 2020-03-31
SHENZHEN GRADUATE SCHOOL TSINGHUA UNIV
View PDF5 Cites 3 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0003] 1) In the early stage, a large amount of user background information that has nothing to do with the text itself needs to be collected, and it is often difficult to collect these user background information in actual scenarios;
[0004] 2) It does not really achieve user-level personalization, but selects sentences that are more in line with the selected topic for different topics as the output of personalized summaries
[0005] Because these two problems cannot be overcome, there is a lack of a user-level personalized text summarization method in the prior art
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0028] Such as figure 1 As shown, the present invention provides a user-level personalized text summary generation method, comprising the following steps:
[0029] S1: collecting the basic information of the user and the original data of the user's attention distribution when reading at least one category of text;
[0030] S2: Extracting the user's personalized preference features from the attention distribution original data, the personalized preference features including the user's attention distribution on different parts of speech;
[0031] Specifically, the percentage of the user's attention distribution time on different parts of speech of the same article to the total text reading time is obtained, and the average value of the user's attention distribution percentage on all texts is used as the personalized preference feature.
[0032] S3: Input the user's personalized preference features and the text into the personalized summary extraction module, and output the extr...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention provides a user-level personalized text abstract generation method and system. The method comprises the steps of collecting basic information of a user and attention distribution original data of the user when the user reads at least one category of text; extracting personalized preference characteristics of the user, wherein the personalized preference characteristics comprise attention distribution of the user on different part-of-voice; inputting the personalized preference characteristics and texts of the user into a personalized abstract extraction module, and outputting thepersonalized preference characteristics and texts as extracted personalized sentences; combining the personalized sentences into a new text, and inputting the new text into a subsequent abstract generation model to generate a personalized text abstract; obtaining basic information of a new user, calculating the similarity with an existing user, selecting at least two most similar users, and obtaining an average value of personalized preference characteristics as initial personalized preference characteristics of the new user; and obtaining a personalized text abstract of the new user according to the initialized personalized preference characteristics. The personalized text abstract is obtained by matching the basic information of the new user, and the method is simple and accurate.
Description
technical field [0001] The present invention relates to the technical field of text information mining, in particular to a user-level personalized text abstract generation method and system. Background technique [0002] Existing text summarization technologies are mainly non-personalized text summarization types, which are completely the same for the summarization generated by different users of the same text, and cannot generate personalized text summaries for different users. However, the existing personalized text summarization technology mainly has two problems: [0003] 1) In the early stage, a large amount of user background information that has nothing to do with the text itself needs to be collected, but it is often difficult to collect these user background information in actual scenarios; [0004] 2) It does not really achieve user-level personalization, but selects sentences that are more in line with the selected topic for different topics as the output of pers...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06F16/34G06F16/335G06F16/9536
CPCG06F16/345G06F16/335G06F16/9536
Inventor 王智易坤孙立峰
Owner SHENZHEN GRADUATE SCHOOL TSINGHUA UNIV