CN106250513B - Event modeling-based event personalized classification method and system - Google Patents
Event modeling-based event personalized classification method and system Download PDFInfo
- Publication number
- CN106250513B CN106250513B CN201610630479.6A CN201610630479A CN106250513B CN 106250513 B CN106250513 B CN 106250513B CN 201610630479 A CN201610630479 A CN 201610630479A CN 106250513 B CN106250513 B CN 106250513B
- Authority
- CN
- China
- Prior art keywords
- event
- classification
- library
- theme
- news
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
- G06F16/355—Class or cluster creation or modification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/955—Retrieval from the web using information identifiers, e.g. uniform resource locators [URL]
- G06F16/9562—Bookmark management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
Abstract
The invention discloses an event personalized classification method and system based on event modeling, wherein the method comprises the following steps of 1, performing data crawling on daily news, corresponding microblogs and comments on all big forums, and establishing a news library; 2. deeply mining data in a news library, establishing a corresponding event model, generating a plurality of labels for each event, and then storing event information in the event library; 3. different classification conditions and keywords of each type required by different users are guided to be input through the client; 4. based on the event labels and other attributes, the events in the event library are classified into different categories according to category matching and category similarity, and are displayed on a user platform in an intuitive mode. The system can make full use of the internet which is rapidly developed at present and a large amount of emerging news data, realize the personalized classification of news event data according to different requirements, can be conveniently and fully butted with various news platforms, and provides personalized information service for various platforms.
Description
Technical Field
The invention relates to the technical field of social network information processing, in particular to the technical field of social network event classification, and particularly relates to an event personalized classification method and system based on event modeling.
Background
With the further popularization of networks and the rapid development of social media, various public opinion information and event dissemination ways are increasing and faster, the information amount generally shows an explosive trend, and if the information cannot be classified well, the user is difficult to be helped to efficiently acquire the information desired by the user.
Conventional event classification techniques are often based on determining categories, i.e., into which categories events are classified are predetermined, such that the classification of an event library is targeted; in addition, in the conventional classification method, an event is only classified into a specific class, that is, the class attribute of an event is unique. With the increasing amount of internet data, news data of a website often comes from many source stations, and the classification criteria of the source stations cannot meet the classification requirements of the destination websites (for example, an event of the source station is classified into a technology class, and the destination websites do not have the technology class). Therefore, how to utilize the uniform event library to perform personalized classification on events according to the classification standards of different platforms so as to facilitate the display of event information on different platforms is a problem frequently encountered in information acquisition and event analysis under the background of big data.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides an event personalized classification method and system based on event modeling.
The purpose of the invention is realized by the following technical scheme: an event personalized classification method based on event modeling comprises the following steps:
s1: acquiring dynamic information of corresponding time, crawling news occurring in near x days, corresponding microblogs and comments on each big forum, and establishing a news library;
s2: deeply mining data acquired from each large platform, establishing an event model, and storing the event model in an event library;
s3: the events are classified according to different classification requirements.
The step S2 of deeply mining the data, establishing an event model, and storing the event model in an event library includes the following substeps:
s21: extracting words in the event corpus, storing the obtained words into a word bank after segmenting the corpus and removing stop words, and updating the word bank;
s22: acquiring a theme classification library of corresponding identification events, establishing the theme classification library according to the number of words of the corpus and the data characteristics acquired by 1/10 of the order of magnitude of all words in a time window, and if the order of magnitude of a word is unchanged, the theme classification library is unchanged for the next time window, and if the order of magnitude of the word is changed greatly, the calculation is carried out again, so that the classification accuracy is improved;
s23: dividing the topics of each event, calculating the similarity between the events and each topic by using the text features of the events, and storing the topics with the similarity larger than a first threshold;
s24: acquiring labels of corresponding events, and labeling the events by using the characteristics and the theme of the existing data, wherein a plurality of labels may exist in one event, and the labels with the threshold value larger than a specific value and the probability of each label need to be stored in an event library;
s25: and establishing a corresponding event model, and establishing a model classified according to the subject or the label of the event based on the subject of the event and the label of the event.
The establishment of the subject classification library adopts an LDA algorithm.
And (3) labeling the event by adopting a TF-IWF algorithm or a TF-IDF algorithm, wherein the TF-IDF algorithm formula is as follows:
description of the drawings:
representing the proportion of the number of times of the word appearing in the article to the total number of words in the article, wherein M is the number of times of the word appearing, M is the number of total words in the article, N is the total number of the articles, and N is the number of the articles containing the word;
the TF-IWF algorithm formula is as follows:
description of the drawings:
the number of times of the word appearing in the article is represented as the proportion of the total number of words in the article, M is the number of times of the word appearing, M is the number of total words in the article, K is the total number of words in all articles, and K is the number of times of the word appearing in all articles.
The step S3 of classifying the event includes the following sub-steps:
s31: obtaining a classification requirement;
s32: firstly, calculating the similarity degree of a theme and the theme in a corresponding theme classification library according to an event word vector or Jaccard similarity calculation method, and if the matching degree of the event theme exceeds a preset threshold value Q, dividing the theme; similarly, calculating the label category of the event; if the similarity degree between the theme of the event and the themes in all theme libraries does not exceed a set threshold value Q, finding out K relatively most similar themes as the theme category of the event; and similarly, obtaining the label of the event, wherein K is a preset numerical value.
The calculation formula of the Jaccard similarity calculation method is as follows:
the classification requirements include which classes to classify, keywords for each class, and whether an event can belong to multiple classes.
An event personalized classification system based on event modeling, comprising:
the data crawling module is mainly used for crawling the information of each news website, related forums, microblogs and related comment information thereof;
the news library stores various news and comments crawled by the data crawling module;
the event analysis module is mainly used for carrying out data mining and event modeling on information in a news library;
the event library stores various event information generated by the event analysis module;
the classification module is mainly used for acquiring a classification standard and classifying according to the classification standard and a classification rule;
the user front-end module provides an interface for a user to facilitate the user to enter a classification method, and the user is usually a manager of a certain website or an information platform;
and the user back-end module is mainly used for calling the classification function of the classification module, obtaining a classification result according to the classification standard of the user, and storing the classification result and then providing the classification result for a specific website to call.
The invention has the beneficial effects that: the invention provides an event personalized classification method and system based on event modeling, which utilizes the characteristics and themes of events to generate a series of event labels, and calculates the similarity between the event labels and the classification according to the classification requirements of users, thereby carrying out personalized classification on the events and ensuring that the events of an event library can meet the requirements of classified display on different platforms; the method comprises the steps of obtaining internet large-scale real-time event data for analysis, wherein the scale of events is determined by the number of crawled source stations; the event in the event library is marked by multiple labels, and each label has a specific weight, so that the category attribution of the event can be conveniently calculated; the events can be flexibly and individually classified according to the requirements of users, and are not limited to a specific classification mode.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
fig. 2 is a block diagram of the system of the present invention.
Detailed Description
The technical solutions of the present invention are further described in detail below with reference to the accompanying drawings, but the scope of the present invention is not limited to the following.
An event personalized classification method based on event modeling comprises the following steps:
s1: acquiring dynamic information of corresponding time, crawling news occurring in near x days, corresponding microblogs and comments on each big forum, and establishing a news library;
s2: deeply mining data acquired from each large platform, establishing an event model, and storing the event model in an event library;
s3: the events are classified according to different classification requirements.
The step S2 of deeply mining the data, establishing an event model, and storing the event model in an event library includes the following substeps:
s21: extracting words in the event corpus, storing the obtained words into a word bank after segmenting the corpus and removing stop words, and updating the word bank;
s22: obtaining a theme classification library of corresponding identification events, establishing the theme classification library according to the number of words of the corpus and the data characteristics obtained by 1/10 of the order of magnitude of all words in a time window, and if the order of magnitude of the words is unchanged, the theme classification library is unchanged for the next time window, and if the order of magnitude of the words is changed greatly, the theme classification library is recalculated, so that the classification accuracy is improved, wherein the time window is determined according to actual needs and is generally determined as a week;
s23: dividing the topics of each event, calculating the similarity between the event and each topic by using the text features of the event, storing the topics with the similarity greater than a first threshold, and determining the threshold by the first threshold according to the requirement on the similarity;
s24: acquiring labels of corresponding events, and labeling the events by using the characteristics and the theme of the existing data, wherein a plurality of labels may exist in one event, and the labels with the threshold value larger than a specific value and the probability of each label need to be stored in an event library;
s25: and establishing a corresponding event model, and establishing a model classified according to the subject or the label of the event based on the subject of the event and the label of the event.
The establishment of the subject classification library adopts an LDA algorithm.
And (3) labeling the event by adopting a TF-IWF algorithm or a TF-IDF algorithm, wherein the TF-IDF algorithm formula is as follows:
description of the drawings:
representing the proportion of the number of times of the word appearing in the article to the total number of words in the article, wherein M is the number of times of the word appearing, M is the number of total words in the article, N is the total number of the articles, and N is the number of the articles containing the word;
the TF-IWF algorithm formula is as follows:
description of the drawings:
the number of times of the word appearing in the article is represented as the proportion of the total number of words in the article, M is the number of times of the word appearing, M is the number of total words in the article, K is the total number of words in all articles, and K is the number of times of the word appearing in all articles.
The step S3 of classifying the event includes the following sub-steps:
s31: obtaining a classification requirement;
s32: firstly, calculating the similarity degree of a theme and the theme in a corresponding theme classification library according to an event word vector or Jaccard similarity calculation method, and if the matching degree of the event theme exceeds a preset threshold value Q, dividing the theme; similarly, calculating the label category of the event; if the similarity degree between the theme of the event and the themes in all theme libraries does not exceed a set threshold value Q, finding out K relatively most similar themes as the theme category of the event; and similarly, obtaining the label of the event, wherein K is a preset numerical value.
The calculation formula of the Jaccard similarity calculation method is as follows:
the classification requirements include which classes to classify, keywords for each class, and whether an event can belong to multiple classes.
As shown in fig. 2, an event personalized classification system based on event modeling includes:
the data crawling module is mainly used for crawling the information of each news website, related forums, microblogs and related comment information thereof;
the news library stores various news and comments crawled by the data crawling module;
the event analysis module is mainly used for carrying out data mining and event modeling on information in a news library;
the event library stores various event information generated by the event analysis module;
the classification module is mainly used for acquiring a classification standard and classifying according to the classification standard and a classification rule;
the user front-end module provides an interface for a user to facilitate the user to enter a classification method, and the user is usually a manager of a certain website or an information platform;
and the user back-end module is mainly used for calling the classification function of the classification module, obtaining a classification result according to the classification standard of the user, and storing the classification result and then providing the classification result for a specific website to call.
The present invention is further described in detail with reference to fig. 1, in this embodiment, a browser is used as a carrier, and the main process is as follows:
step one, performing data crawling on news which occur every day and comments on large forums such as microblogs and the like, and establishing a news library. The main websites include: crawling of news and related comments of various large websites such as internet surfing, Tencent, people's network, Skyline, 91 and the like.
And step two, performing deep mining on the data crawled from the big news and forum platforms, establishing corresponding event models, and storing the event models in an event library. The method mainly comprises the following four steps: data preprocessing, theme extraction, label extraction and event clustering. The modeled event model and the analysis results are then stored in an event repository.
And step three, designing a user side page to facilitate the user to input the classification standard.
And step four, dividing the events in the event library into five categories of society, entertainment, finance, sports, international and the like according to the requirements of the user, wherein each category has a plurality of keywords for explaining the classification basis of the category. According to the classification of the category, if one of the 5 labels of a certain event exists, the event is classified into the category, if not, the similarity between the label and the category key word is calculated, and the event is classified into the category with the highest similarity.
Claims (5)
1. An event personalized classification method based on event modeling is characterized by comprising the following steps:
s1: acquiring dynamic information of corresponding time, crawling news occurring in near x days, corresponding microblogs and comments on each big forum, and establishing a news library;
s2: deeply mining data acquired from each large platform, establishing an event model, and storing the event model in an event library;
the step S2 of deeply mining the data, establishing an event model, and storing the event model in an event library includes the following substeps:
s21: extracting words in the event corpus, storing the obtained words into a word bank after segmenting the corpus and removing stop words, and updating the word bank;
s22: acquiring a theme classification library of corresponding identification events, establishing the theme classification library according to the number of words of the corpus and the data characteristics acquired by 1/10 of the order of magnitude of all words in a time window, and if the order of magnitude of a word is unchanged, the theme classification library is unchanged for the next time window, and if the order of magnitude of the word is changed greatly, the calculation is carried out again, so that the classification accuracy is improved;
s23: dividing the topics of each event, calculating the similarity between the events and each topic by using the text characteristics of the events, and storing the topics with the similarity larger than a threshold value;
s24: acquiring labels of corresponding events, and labeling the events by using the characteristics and the theme of the existing data, wherein a plurality of labels may exist in one event, and the labels with the threshold value larger than a specific value and the probability of each label need to be stored in an event library;
s25: establishing a corresponding event model, and establishing a model classified according to the subject or the label of the event based on the subject of the event and the label of the event;
s3: classifying the events according to different classification requirements;
the step S3 of classifying the event includes the following sub-steps:
s31: obtaining classification requirements, wherein the classification requirements comprise classes to which the events are classified, keywords of each class and whether an event belongs to a plurality of classes;
s32: firstly, calculating the similarity degree of a theme and the theme in a corresponding theme classification library according to an event word vector or Jaccard similarity calculation method, and if the matching degree of the event theme exceeds a preset threshold value Q, dividing the theme; similarly, calculating the label category of the event; if the similarity degree between the theme of the event and the themes in all theme libraries does not exceed a set threshold value Q, finding out K relatively most similar themes as the theme category of the event; and similarly, obtaining the label of the event, wherein K is a preset numerical value.
2. The event personalized classification method based on event modeling according to claim 1, characterized in that: the establishment of the subject classification library adopts an LDA algorithm.
3. The event personalized classification method based on event modeling according to claim 1, characterized in that: and (3) labeling the event by adopting a TF-IWF algorithm or a TF-IDF algorithm, wherein the TF-IDF algorithm formula is as follows:
description of the drawings:
representing the proportion of the number of times of the word appearing in the article to the total number of words in the article, wherein M is the number of times of the word appearing, M is the number of total words in the article, N is the total number of the articles, and N is the number of the articles containing the word;
the TF-IWF algorithm formula is as follows:
description of the drawings:
the number of times of the word appearing in the article is represented as the proportion of the total number of words in the article, M is the number of times of the word appearing, M is the number of total words in the article, K is the total number of words in all articles, and K is the number of times of the word appearing in all articles.
4. The event personalized classification method based on event modeling according to claim 1, characterized in that: the calculation formula of the Jaccard similarity calculation method is as follows:
5. an event-modeling based event personalization classification system of an event-modeling based event personalization classification method according to any one of claims 1 to 4, characterized in that it comprises:
the data crawling module is mainly used for crawling the information of each news website, related forums, microblogs and related comment information thereof;
the news library stores various news and comments crawled by the data crawling module;
the event analysis module is mainly used for carrying out data mining and event modeling on information in a news library;
the event library stores various event information generated by the event analysis module;
the classification module is mainly used for acquiring a classification standard and classifying according to the classification standard and a classification rule;
the user front-end module provides an interface for a user to facilitate the user to enter a classification method, and the user is usually a manager of a certain website or an information platform;
and the user back-end module is mainly used for calling the classification function of the classification module, obtaining a classification result according to the classification standard of the user, and storing the classification result and then providing the classification result for a specific website to call.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610630479.6A CN106250513B (en) | 2016-08-02 | 2016-08-02 | Event modeling-based event personalized classification method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610630479.6A CN106250513B (en) | 2016-08-02 | 2016-08-02 | Event modeling-based event personalized classification method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106250513A CN106250513A (en) | 2016-12-21 |
| CN106250513B true CN106250513B (en) | 2021-04-23 |
Family
ID=57606479
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610630479.6A Active CN106250513B (en) | 2016-08-02 | 2016-08-02 | Event modeling-based event personalized classification method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106250513B (en) |
Families Citing this family (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108320255B (en) * | 2017-01-16 | 2022-06-21 | 软通动力信息技术(集团)股份有限公司 | Information processing method and device |
| CN107797983A (en) * | 2017-04-07 | 2018-03-13 | 平安科技（深圳）有限公司 | Microblog data processing method, device, computer equipment and storage medium |
| CN107231570A (en) * | 2017-06-13 | 2017-10-03 | 中国传媒大学 | News data content characteristic obtains system and application system |
| CN107562785A (en) * | 2017-07-26 | 2018-01-09 | 深圳市赛亿科技开发有限公司 | A kind of reading content collating sort method and apparatus |
| CN107741929A (en) * | 2017-10-18 | 2018-02-27 | 网智天元科技集团股份有限公司 | The analysis of public opinion method and device |
| CN108536673B (en) * | 2018-03-16 | 2022-06-21 | 数库（上海）科技有限公司 | News event extraction method and device |
| CN108763189B (en) * | 2018-04-12 | 2022-03-25 | 武汉斗鱼网络科技有限公司 | Live broadcast room content label weight calculation method and device and electronic equipment |
| CN109409529B (en) * | 2018-09-13 | 2020-12-08 | 北京中科闻歌科技股份有限公司 | Event cognitive analysis method, system and storage medium |
| CN109241438B (en) * | 2018-09-27 | 2022-06-24 | 国家计算机网络与信息安全管理中心 | Element-based cross-channel hot event discovery method and device and storage medium |
| CN111078867A (en) * | 2018-10-19 | 2020-04-28 | 北京国双科技有限公司 | Text classification method and device |
| CN109299273B (en) * | 2018-11-02 | 2020-06-23 | 广州语义科技有限公司 | Multi-source multi-label text classification method and system based on improved seq2seq model |
| CN110457468B (en) * | 2019-07-05 | 2022-08-23 | 武楚荷 | Event classification method and device and storage device |
| CN111428127B (en) * | 2020-01-21 | 2023-08-11 | 江西财经大学 | Personalized event recommendation method and system integrating theme matching and bidirectional preference |
| CN111461266B (en) * | 2020-06-18 | 2020-12-01 | 爱保科技有限公司 | Vehicle damage assessment abnormity identification method, device, server and storage medium |
| CN112084448A (en) * | 2020-08-31 | 2020-12-15 | 北京金堤征信服务有限公司 | Similar information processing method and device |
| CN113435199B (en) * | 2021-07-18 | 2023-05-26 | 谢勇 | Storage and reading interference method and system for character corresponding culture |
| CN113656697B (en) * | 2021-08-24 | 2023-12-12 | 北京字跳网络技术有限公司 | Object recommendation method, device, electronic equipment and storage medium |
| CN113821754A (en) * | 2021-09-18 | 2021-12-21 | 上海观安信息技术股份有限公司 | Sensitive data interface crawler identification method and device |
| CN115329903B (en) * | 2022-10-12 | 2023-05-30 | 福建美舫时代科技有限公司 | Spatial data integration method and system applied to digital twin city |
| CN116634230B (en) * | 2023-05-24 | 2024-03-22 | 天津大学 | Throwing method based on multi-channel new media hot event spreading effect analysis |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102495872A (en) * | 2011-11-30 | 2012-06-13 | 中国科学技术大学 | Method and device for conducting personalized news recommendation to mobile device users |
| US20120290518A1 (en) * | 2011-03-29 | 2012-11-15 | Manyworlds, Inc. | Integrated search and adaptive discovery system and method |
| CN102831234A (en) * | 2012-08-31 | 2012-12-19 | 北京邮电大学 | Personalized news recommendation device and method based on news content and theme feature |
| CN104217038A (en) * | 2014-09-30 | 2014-12-17 | 中国科学技术大学 | Knowledge network building method for financial news |
| CN105045812A (en) * | 2015-06-18 | 2015-11-11 | 上海高欣计算机系统有限公司 | Text topic classification method and system |
| CN105550365A (en) * | 2016-01-15 | 2016-05-04 | 中国科学院自动化研究所 | Visualization analysis system based on text topic model |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104679875B (en) * | 2015-03-10 | 2017-12-15 | 杭州凡闻科技有限公司 | A kind of information data classification method based on digital newspaper |
-
2016
- 2016-08-02 CN CN201610630479.6A patent/CN106250513B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120290518A1 (en) * | 2011-03-29 | 2012-11-15 | Manyworlds, Inc. | Integrated search and adaptive discovery system and method |
| CN102495872A (en) * | 2011-11-30 | 2012-06-13 | 中国科学技术大学 | Method and device for conducting personalized news recommendation to mobile device users |
| CN102831234A (en) * | 2012-08-31 | 2012-12-19 | 北京邮电大学 | Personalized news recommendation device and method based on news content and theme feature |
| CN104217038A (en) * | 2014-09-30 | 2014-12-17 | 中国科学技术大学 | Knowledge network building method for financial news |
| CN105045812A (en) * | 2015-06-18 | 2015-11-11 | 上海高欣计算机系统有限公司 | Text topic classification method and system |
| CN105550365A (en) * | 2016-01-15 | 2016-05-04 | 中国科学院自动化研究所 | Visualization analysis system based on text topic model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106250513A (en) | 2016-12-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106250513B (en) | Event modeling-based event personalized classification method and system | |
| CN107204184B (en) | Audio recognition method and system | |
| CN109325165B (en) | Network public opinion analysis method, device and storage medium | |
| CN109145216A (en) | Network public-opinion monitoring method, device and storage medium | |
| CN107220386A (en) | Information-pushing method and device | |
| WO2019041521A1 (en) | Apparatus and method for extracting user keyword, and computer-readable storage medium | |
| US9436768B2 (en) | System and method for pushing and distributing promotion content | |
| CN106874258B (en) | Text similarity calculation method and system based on Chinese character attribute vector representation | |
| CN103020159A (en) | Method and device for news presentation facing events | |
| CN101593200A (en) | Chinese Web page classification method based on the keyword frequency analysis | |
| CN107944032B (en) | Method and apparatus for generating information | |
| CN108959329B (en) | Text classification method, device, medium and equipment | |
| CN112632278A (en) | Labeling method, device, equipment and storage medium based on multi-label classification | |
| WO2021184640A1 (en) | Sparse matrix-based product pushing method and apparatus, computer device, and medium | |
| CN110321553A (en) | Short text subject identifying method, device and computer readable storage medium | |
| CN106354818A (en) | Dynamic user attribute extraction method based on social media | |
| Chumwatana | Using sentiment analysis technique for analyzing Thai customer satisfaction from social media | |
| CN113282754A (en) | Public opinion detection method, device, equipment and storage medium for news events | |
| CN110019820A (en) | Main suit and present illness history symptom Timing Coincidence Detection method in a kind of case history | |
| CN112818230A (en) | Content recommendation method and device, electronic equipment and storage medium | |
| CN107908749B (en) | Character retrieval system and method based on search engine | |
| CN104240107B (en) | Community data screening system and method thereof | |
| CN111651559B (en) | Social network user relation extraction method based on event extraction | |
| CN108959289B (en) | Website category acquisition method and device | |
| CN106407473B (en) | event similarity modeling-based method and system for acquiring event context |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |