CN107103093B - Short text recommendation method and device based on user behavior and emotion analysis - Google Patents

Short text recommendation method and device based on user behavior and emotion analysis Download PDF

Info

Publication number
CN107103093B
CN107103093B CN201710345317.2A CN201710345317A CN107103093B CN 107103093 B CN107103093 B CN 107103093B CN 201710345317 A CN201710345317 A CN 201710345317A CN 107103093 B CN107103093 B CN 107103093B
Authority
CN
China
Prior art keywords
user
emotion
short
text
short text
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.)
Expired - Fee Related
Application number
CN201710345317.2A
Other languages
Chinese (zh)
Other versions
CN107103093A (en
Inventor
刘金硕
李瞧
邓娟
刘必为
陈煜森
杨广益
李晨曦
李扬眉
房金城
谈聪
陈凯
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan University WHU
Original Assignee
Wuhan University WHU
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Wuhan University WHU filed Critical Wuhan University WHU
Priority to CN201710345317.2A priority Critical patent/CN107103093B/en
Publication of CN107103093A publication Critical patent/CN107103093A/en
Application granted granted Critical
Publication of CN107103093B publication Critical patent/CN107103093B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/3331—Query processing
    • G06F16/3332—Query translation
    • G06F16/3334—Selection or weighting of terms from queries, including natural language queries
    • 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
    • G06F16/33—Querying
    • G06F16/3331—Query processing
    • G06F16/334—Query execution
    • G06F16/3344—Query execution using natural language analysis
    • 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
    • G06F16/33—Querying
    • G06F16/335—Filtering based on additional data, e.g. user or group profiles
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/30—Semantic analysis

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention relates to a short text recommendation method and device, belongs to the technical field of information processing, and particularly relates to a short text recommendation method and device based on user behavior and emotion analysis. According to the method, the historical behaviors of the user and the historical short text emotion of the user are combined, the short texts of related categories are matched with the user through negative emotion correlation analysis, and the related short texts are recommended.

Description

Short text recommendation method and device based on user behavior and emotion analysis
Technical Field
The invention relates to a short text recommendation method and device, belongs to the technical field of information processing, and particularly relates to a short text recommendation method and device based on user behavior and emotion analysis.
Background
With the development of internet technology and social networks, networks have become a main source for people to obtain information, people spend more and more time on the social networks, and the situation is more prone to express emotions on the social networks. The traditional social network recommendation method cannot meet the discovery requirement of the user on the information, because sometimes the user cannot accurately describe the current requirement.
The recommendation algorithm is used for guessing user preferences according to historical behaviors of users, personal information and the like, and recommending contents or articles which the users may be interested in. In the prior art, recommendation algorithms are mainly divided into two types: the first is based on content filtering, and the recommendation system based on content filtering realizes a recommendation function by comparing the similarity between commodities rather than the similarity between users; the second is based on collaborative filtering, and the e-commerce recommendation system based on collaborative filtering does not analyze the similarity between commodities but learns the similarity of behaviors between the target user and the historical user without depending on the characteristics of the commodities, so as to generate a recommendation result according to the behaviors of similar historical users.
However, in the above prior art, the content-based filtering can take into account the similarity of short texts and cannot take into account the timeliness of the short texts, so the recommendation effect is not ideal; based on collaborative filtering, recommendation is required in view of visiting users, and only short texts of visited people are recommended for short text recommendation with higher timeliness requirement, so that some overdue popular short texts are generated in the collaborative filtering. Meanwhile, the technology does not consider the emotion problem of the user when the short text is recommended for the user, and the short text which accords with the mood of the user and can actively and positively guide the emotion of the user cannot be recommended.
Disclosure of Invention
Aiming at the problem that a large amount of short texts in the social network can not be effectively and reasonably recommended to related users and can not guide the emotion of the users in the forward direction, the method combines the historical behaviors of the users and the historical short texts of the users, matches the short texts and the users of related categories through negative emotion correlation analysis, and recommends the related short texts.
The technical problem of the invention is mainly solved by the following technical scheme:
a short text recommendation method based on user behavior and emotion analysis comprises the following steps:
step 1, generating a user characteristic vector based on the acquired user historical data;
step 2, acquiring a reference user with similar historical behavior to a target user as a similar user group of the target user based on the user characteristic vector;
step 3, calculating short text emotional characteristic vector E recently published by the target user1；
And 4, recommending the short text to the user based on the emotion vector of the short text recently published by the user.
Preferably, in the short text recommendation method based on user behavior and emotion analysis, the step 1 specifically includes:
step 1.1, calculating the frequency of publishing the original short text by the user within a period of time, and recording the frequency as FO;
and step 1.2, calculating the frequency of publishing and forwarding short texts in a period of time of the user, and recording the frequency as FT.
Step 1.3, analyzing the short text and self-introduction published by the user by using an LDA topic model, and acquiring topic characteristics which are marked as T;
and 1.4, constructing a user feature vector based on FO, FT and T.
Preferably, in the short text recommendation method based on user behavior and emotion analysis, the step 2 specifically includes:
step 2.1, calculating the similarity of published behaviors of users
Figure BDA0001296305120000031
Step 2.2, calculating the theme similarity rho of the short textT＝Num(T1∩T)；
Step 2.3, based on the formula rhoH＝w1ρT+w2ρuAcquiring user behavior similarity, sequencing the reference users in a descending order according to the user behavior similarity, and taking the first N reference users as a similar user group of a target user;
in the formula, the frequency FT of the original short text of the target user1Forwarding short text frequency FO of target subscriber1Target userSubject feature T of1，w1,w2And respectively giving out the similarity of the published behaviors of the user and the weight of the similarity of the published behaviors of the user.
Preferably, in the short text recommendation method based on user behavior and emotion analysis, the step 3 specifically includes:
step 3.1, performing word segmentation on the short text recently published by the target user;
step 3.2, counting the occurrence frequency of words with different parts of speech, filling the occurrence frequency of words with different parts of speech into an array in sequence, and recording as the emotion vector E of the short text1；
Step 3.3, based on the formula e1Computing an emotion vector E ═ x-motion-y-motion1The sentiment value of (2); wherein, motion is the corresponding weight of each positive emotion word, motion is the corresponding weight of each negative emotion word, and x is an emotion vector E1The score corresponding to each positive emotion word is shown in the specification, and y is an emotion vector E1The score corresponding to the negative emotion word in (1).
Preferably, in the short text recommendation method based on user behavior and emotion analysis, the step 4 specifically includes:
step 4.1, selecting the short texts recently published by the similar user groups into a text set to be recommended, and acquiring the emotion similarity rho between each text to be recommended in the text set to be recommended and the short text recently published by the target user based on the following formula:
Figure BDA0001296305120000032
in the formula (I), the compound is shown in the specification,
Figure BDA0001296305120000041
issuing short text for the emotion vector E of the text to be recommended and the emotion vector E of the target user recently1Standard deviation of (d);
and 4.2, selecting a plurality of texts to be recommended from the text set to be recommended as candidate texts according to the emotion similarity rho.
Preferably, in the short text recommendation method based on user behavior and emotion analysis, in step 4.2,
when e is1<When 0, comparing the short texts in the text set to be recommended with the short texts recently published by the target user one by one, and finally obtaining three short texts with rho closest to-1 as candidate recommendations; when e is1>And when 0, comparing the short texts in the text set to be recommended with the short texts recently published by the target user one by one, and finally obtaining three short texts which are the most approximate to 1 and are used as candidate recommendations.
Preferably, in the short text recommendation method based on user behavior and emotion analysis, the short text with the highest praise number is selected as the recommended short text from the candidate recommended short texts.
A short text recommendation apparatus based on user behavior and emotion analysis, comprising:
the characteristic extraction module generates a user characteristic vector based on the acquired user historical data;
the group division module is used for acquiring a reference user with a history behavior similar to that of a target user as a similar user group of the target user based on the user characteristic vector;
the emotion analysis module is used for calculating short text emotion characteristic vector E recently published by the target user1；
And the text recommendation module is used for recommending the short text to the user based on the emotion vector of the short text recently published by the user.
Preferably, the short text recommendation device based on user behavior and emotion analysis includes:
the original frequency calculating unit is used for calculating the frequency of publishing the original short text by the user within a period of time and recording the frequency as FO;
and the forwarding frequency calculation unit is used for calculating the frequency of issuing and forwarding short texts within a period of time of the user and recording the frequency as FT.
The topic feature calculation unit analyzes the short texts and self-introduction published by the user by using an LDA topic model to obtain topic features which are marked as T;
and a feature vector construction unit for constructing the user feature vector based on F0, FT and T.
Preferably, the short text recommendation device based on user behavior and emotion analysis includes:
the text word segmentation unit is used for segmenting short texts recently published by the target user;
a vector construction unit for counting the occurrence frequency of words with different parts of speech, filling the occurrence frequency of words with different parts of speech into an array in sequence, and recording as the emotion vector E of the short text1；
Sentiment scoring unit based on formula e1Computing an emotion vector E ═ x-motion-y-motion1The sentiment value of (2); wherein, motion is the corresponding weight of each positive emotion word, motion is the corresponding weight of each negative emotion word, and x is an emotion vector E1The score corresponding to each positive emotion word is shown in the specification, and y is an emotion vector E1The score corresponding to the negative emotion word in (1).
Therefore, the invention has the following advantages:
(1) the timeliness is strong, and interested short texts with strong timeliness can be recommended to the user in time instead of overdue popular short texts;
(2) the accuracy is high, and the short text which accords with the mood of the user and actively and positively guides the emotion of the user can be recommended to the user;
(3) the method has strong purposiveness and can achieve the effect of emotional intervention by effectively recommending the short text to the user.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention.
Fig. 2 is a flow chart of the present invention for obtaining similar user groups.
FIG. 3 is a flow chart of analyzing user emotion recommendation short text in the present invention.
Detailed Description
The technical scheme of the invention is further specifically described by the following embodiments and the accompanying drawings.
Example (b):
aiming at the problem that a large amount of short texts in the social network can not be effectively and reasonably recommended to related users and can not guide the emotions of the users in the positive direction, the method and the system match the short texts and the users of related categories by analyzing the historical behaviors of the users and the emotions of the historical short texts of the users and analyzing the negative emotions and recommend the related short texts.
The invention provides a short text recommendation method based on user behavior and emotion analysis, the implementation flow of which is shown in figure 1, and the method comprises the following steps:
firstly, historical data of a relevant user, including user self introduction, recently published text content, original text quantity, forwarded text quantity and published text praise quantity, is acquired and stored in a database.
The user behavior is analyzed, user clustering is performed, a user group similar to the target user is obtained, and the flow is shown in fig. 2. Marking the frequency of publishing original short texts of a target user within a period of time as FT1Short text forwarding frequency is published within a period of time and is recorded as FO1. Meanwhile, an LDA topic model is used for analyzing the short text and self introduction published by the target user to obtain topic characteristics, and the topic characteristics are recorded as T1. And recording the frequency of publishing the original short text of the user to be compared within a period of time as FT, recording the short text forwarded within a period of time as FO, and recording the theme characteristics obtained by analyzing the published short text and self introduction of the target user by using an LDA theme model as T. Calculating the similarity of the published behaviors of the target user and the user to be compared
Figure BDA0001296305120000061
Calculating the theme similarity rho between the target user and the user to be comparedT＝Num(T1∩ T) user behavior similarity is rhoH＝w1ρT+w2ρu. And taking N users with large user behavior similarity values as a similar user group.
Then, the recent emotion of the target user is analyzed, and a related microblog is recommended to the target user, and the flow is shown in fig. 3. And performing emotion analysis on short texts recently published by the user. Recently published to target users using NLPIR word segmentation toolsDividing words of the short text, counting the occurrence frequency of words with different parts of speech, filling the occurrence frequency of words with different parts of speech into an array in sequence, and recording as an emotion vector E of the short text1. Get E1Has an emotional value of e1The term "motion" refers to x-motion-y-motion, wherein motion is the weight corresponding to each positive emotion word, motion is the weight corresponding to each negative emotion word, x is the parameter corresponding to each positive emotion word in the emotion vector, and y is the parameter corresponding to the negative emotion word in the emotion vector. Selecting all short texts sent by the obtained similar user groups into a text set to be recommended, recording the emotion vector of the text to be recommended as E, and obtaining the emotion similarity
Figure BDA0001296305120000071
When e is1<And 0, comparing the short texts in the text set to be recommended with the short texts recently published by the target user one by one, and finally obtaining three short texts with rho closest to-1 as candidate recommendations. And selecting the short text with the highest praise number from the three candidate recommended short texts as the recommended short text.
The specific embodiments described herein are merely illustrative of the spirit of the invention. Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.

Claims (9)

1. A short text recommendation method based on user behavior and emotion analysis is characterized by comprising the following steps:
step 1, generating a user characteristic vector based on the acquired user historical data;
step 2, obtaining a reference user with a history behavior similar to that of the target user as a similar user group of the target user based on the user feature vector, specifically comprising:
step 2.1, calculating the similarity of published behaviors of users
Figure FDA0002240821110000011
Step 2.2, calculating the theme similarity rho of the short textT＝Num(T1∩T)；
Step 2.3, based on the formula rhoH＝w1ρT+w2ρuAcquiring user behavior similarity, sequencing the reference users in a descending order according to the user behavior similarity, and taking the first N reference users as a similar user group of a target user;
in the formula, the frequency FT of the original short text of the target user1Forwarding short text frequency FO of target subscriber1Topic feature T of the target user1，w1,w2Respectively issuing the weight values of the behavior similarity and the user publication behavior similarity for the user;
step 3, calculating short text emotional characteristic vector E recently published by the target user1；
And 4, recommending the short text to the user based on the emotion vector of the short text recently published by the user.
2. The method for recommending short texts based on user behavior and emotion analysis according to claim 1, wherein the step 1 specifically comprises:
step 1.1, calculating the frequency of publishing the original short text by the user within a period of time, and recording the frequency as FO;
step 1.2, calculating the frequency of publishing and forwarding short texts of a user within a period of time, and recording the frequency as FT;
step 1.3, analyzing the short text and self-introduction published by the user by using an LDA topic model, and acquiring topic characteristics which are marked as T;
and step 1.4, constructing a user feature vector based on F0, FT and T.
3. The method for recommending short texts based on user behavior and emotion analysis according to claim 1, wherein the step 3 specifically comprises:
step 3.1, performing word segmentation on the short text recently published by the target user;
step 3.2, counting different parts of speechThe frequency of the word appearance is filled into the array in sequence and is recorded as the emotion vector E of the short text1；
Step 3.3, based on the formula e1Computing an emotion vector E ═ x-motion-y-motion1The sentiment value of (2); wherein, motion is the corresponding weight of each positive emotion word, motion is the corresponding weight of each negative emotion word, and x is an emotion vector E1The score corresponding to each positive emotion word is shown in the specification, and y is an emotion vector E1The score corresponding to the negative emotion word in (1).
4. The method for recommending short texts based on user behavior and emotion analysis according to claim 3, wherein the step 4 specifically comprises:
step 4.1, selecting the short texts recently published by the similar user groups into a text set to be recommended, and acquiring the emotion similarity rho between each text to be recommended in the text set to be recommended and the short text recently published by the target user based on the following formula:
Figure FDA0002240821110000021
in the formula (I), the compound is shown in the specification,
Figure FDA0002240821110000022
issuing short text for the emotion vector E of the text to be recommended and the emotion vector E of the target user recently1Standard deviation of (d);
and 4.2, selecting a plurality of texts to be recommended from the text set to be recommended as candidate texts according to the emotion similarity rho.
5. The method for recommending short text based on user behavior and emotion analysis as claimed in claim 4, wherein, in step 4.2,
when e is1<When 0, comparing the short texts in the text set to be recommended with the short texts recently published by the target user one by one, and finally obtaining three short texts with rho closest to-1 as candidate recommendations; when e is1>And when 0, comparing the short texts in the text set to be recommended with the short texts recently published by the target user one by one, and finally obtaining three short texts which are the most approximate to 1 and are used as candidate recommendations.
6. The method of claim 5, wherein the short text recommendation method based on the user behavior and emotion analysis is characterized in that the short text with the highest praise number is selected from the candidate recommended short texts as the recommended short text.
7. An apparatus adopting the short text recommendation method based on user behavior and emotion analysis of claim 1, comprising:
the characteristic extraction module generates a user characteristic vector based on the acquired user historical data;
the group division module is used for acquiring a reference user with a history behavior similar to that of a target user as a similar user group of the target user based on the user characteristic vector;
the emotion analysis module is used for calculating short text emotion characteristic vector E recently published by the target user1；
And the text recommendation module is used for recommending the short text to the user based on the emotion vector of the short text recently published by the user.
8. The apparatus for recommending short text based on user behavior and emotion analysis of claim 7, wherein the feature extraction module specifically comprises:
the original frequency calculating unit is used for calculating the frequency of publishing the original short text by the user within a period of time and recording the frequency as FO;
the forwarding frequency calculation unit is used for calculating the frequency of issuing and forwarding short texts within a period of time of the user and recording the frequency as FT;
the topic feature calculation unit analyzes the short texts and self-introduction published by the user by using an LDA topic model to obtain topic features which are marked as T;
and a feature vector construction unit for constructing the user feature vector based on F0, FT and T.
9. The apparatus of claim 7, wherein the emotion analysis module specifically comprises:
the text word segmentation unit is used for segmenting short texts recently published by the target user;
a vector construction unit for counting the occurrence frequency of words with different parts of speech, filling the occurrence frequency of words with different parts of speech into an array in sequence, and recording as the emotion vector E of the short text1；
Sentiment scoring unit based on formula e1Computing an emotion vector E ═ x-motion-y-motion1The sentiment value of (2); wherein, motion is the corresponding weight of each positive emotion word, motion is the corresponding weight of each negative emotion word, and x is an emotion vector E1The score corresponding to each positive emotion word is shown in the specification, and y is an emotion vector E1The score corresponding to the negative emotion word in (1).
CN201710345317.2A 2017-05-16 2017-05-16 Short text recommendation method and device based on user behavior and emotion analysis Expired - Fee Related CN107103093B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710345317.2A CN107103093B (en) 2017-05-16 2017-05-16 Short text recommendation method and device based on user behavior and emotion analysis

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710345317.2A CN107103093B (en) 2017-05-16 2017-05-16 Short text recommendation method and device based on user behavior and emotion analysis

Publications (2)

Publication Number Publication Date
CN107103093A CN107103093A (en) 2017-08-29
CN107103093B true CN107103093B (en) 2020-03-10

Family

ID=59670272

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710345317.2A Expired - Fee Related CN107103093B (en) 2017-05-16 2017-05-16 Short text recommendation method and device based on user behavior and emotion analysis

Country Status (1)

Country Link
CN (1) CN107103093B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110427546A (en) * 2018-04-28 2019-11-08 北京京东尚科信息技术有限公司 A kind of information displaying method and device
CN108920503A (en) * 2018-05-28 2018-11-30 哈尔滨工程大学 A kind of micro- video personalized recommendation algorithm based on social networks degree of belief
CN109271634B (en) * 2018-09-17 2022-07-01 重庆理工大学 Microblog text emotion polarity analysis method based on user emotion tendency perception
CN109446378A (en) * 2018-11-08 2019-03-08 北京奇艺世纪科技有限公司 Information recommendation method, Sentiment orientation determine method and device and electronic equipment
CN110738040B (en) * 2019-10-17 2023-11-28 深圳市比量科技传媒有限公司 Comment expansion method and comment expansion system

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104063399A (en) * 2013-03-22 2014-09-24 杭州金弩信息技术有限公司 Method and system for automatically identifying emotional probability borne by texts
JP2015036856A (en) * 2013-08-12 2015-02-23 大日本印刷株式会社 Information providing apparatus, program, and information providing system

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104063399A (en) * 2013-03-22 2014-09-24 杭州金弩信息技术有限公司 Method and system for automatically identifying emotional probability borne by texts
JP2015036856A (en) * 2013-08-12 2015-02-23 大日本印刷株式会社 Information providing apparatus, program, and information providing system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
基于相似用户情感分析的推荐算法研究;余燕萍;《科技广场》;20160830;第1页第1节到第2页第2节 *

Also Published As

Publication number Publication date
CN107103093A (en) 2017-08-29

Similar Documents

Publication Publication Date Title
CN107103093B (en) Short text recommendation method and device based on user behavior and emotion analysis
CN109933664B (en) Fine-grained emotion analysis improvement method based on emotion word embedding
CN108959603B (en) Personalized recommendation system and method based on deep neural network
CN107608956B (en) Reader emotion distribution prediction algorithm based on CNN-GRNN
CN106503192B (en) Name entity recognition method and device based on artificial intelligence
US9563693B2 (en) Determining sentiments of social posts based on user feedback
CN107357889A (en) A kind of across social platform picture proposed algorithm based on interior perhaps emotion similitude
CN106062730A (en) Systems and methods for actively composing content for use in continuous social communication
US10409915B2 (en) Determining personality profiles based on online social speech
Van Durme Streaming analysis of discourse participants
CN110209954B (en) Group recommendation method based on LDA topic model and deep learning
WO2015021937A1 (en) Method and device for user recommendation
Okazaki et al. How to mine brand Tweets: Procedural guidelines and pretest
CN106126605B (en) Short text classification method based on user portrait
CN108733652B (en) Test method for film evaluation emotion tendency analysis based on machine learning
CN115878841B (en) Short video recommendation method and system based on improved bald eagle search algorithm
KR101712291B1 (en) System for recommending a user-customized famous place based on opinion mining and Method of the Same
Ashraf et al. Personalized news recommendation based on multi-agent framework using social media preferences
CA2895121A1 (en) Systems and methods for analyzing and deriving meaning from large scale data sets
CN114358807A (en) User portrayal method and system based on predictable user characteristic attributes
Tarasova Classification of hate tweets and their reasons using svm
Chen et al. Exploiting aesthetic features in visual contents for movie recommendation
CN108804416B (en) Training method for film evaluation emotion tendency analysis based on machine learning
CN108717450B (en) Analysis algorithm for emotion tendentiousness of film comment
CN111259228A (en) Personalized news recommendation method based on big data deep learning

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20200310

Termination date: 20210516