CN110867250B - Social media self-disabling behavior detection method based on strong robustness feature selection - Google Patents
Social media self-disabling behavior detection method based on strong robustness feature selection Download PDFInfo
- Publication number
- CN110867250B CN110867250B CN201911033392.0A CN201911033392A CN110867250B CN 110867250 B CN110867250 B CN 110867250B CN 201911033392 A CN201911033392 A CN 201911033392A CN 110867250 B CN110867250 B CN 110867250B
- Authority
- CN
- China
- Prior art keywords
- post
- self
- posts
- social media
- 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.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 39
- 238000000034 method Methods 0.000 claims description 36
- 238000012549 training Methods 0.000 claims description 31
- 239000011159 matrix material Substances 0.000 claims description 30
- 239000013598 vector Substances 0.000 claims description 22
- 230000008451 emotion Effects 0.000 claims description 12
- 238000000605 extraction Methods 0.000 claims description 11
- 238000004458 analytical method Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 10
- 206010022524 Intentional self-injury Diseases 0.000 claims description 6
- 208000005560 Self Mutilation Diseases 0.000 claims description 6
- 230000009193 crawling Effects 0.000 claims description 5
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 238000013480 data collection Methods 0.000 claims description 3
- 238000002372 labelling Methods 0.000 claims description 3
- 230000007935 neutral effect Effects 0.000 claims description 3
- 238000004445 quantitative analysis Methods 0.000 claims description 3
- 238000010187 selection method Methods 0.000 claims description 2
- 230000002123 temporal effect Effects 0.000 claims description 2
- 230000002996 emotional effect Effects 0.000 claims 3
- 235000010627 Phaseolus vulgaris Nutrition 0.000 claims 1
- 244000046052 Phaseolus vulgaris Species 0.000 claims 1
- 229920002472 Starch Polymers 0.000 claims 1
- 235000019698 starch Nutrition 0.000 claims 1
- 239000008107 starch Substances 0.000 claims 1
- 230000006399 behavior Effects 0.000 description 21
- 230000002159 abnormal effect Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000007418 data mining Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 206010010144 Completed suicide Diseases 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 201000011243 gastrointestinal stromal tumor Diseases 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 210000000653 nervous system Anatomy 0.000 description 1
- 230000005180 public health Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 229920006395 saturated elastomer Polymers 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Public Health (AREA)
- Medical Informatics (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Biomedical Technology (AREA)
- Multimedia (AREA)
- Pathology (AREA)
- General Engineering & Computer Science (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a social media self-disabling behavior detection method based on strong robustness characteristic selection, which comprises the following steps of 1) obtaining multi-dimensional heterogeneous information from a network social media website; 2) Extracting the characteristics of data from four aspects of text, user, time and picture to construct a self-residual content data set and a normal content data set; 3) Constructing a supervised self-residual detection model selected based on the strong robustness characteristic through a loss function of the l _2,1 norm and a regularization term; 4) And (4) extracting the characteristics of the target data to be detected, and performing self-residual detection by using the constructed detection model. Compared with the traditional self-residual detection, the self-residual detection method for the social media can be used for contacting with the self-residual subject more widely, exploring the behavior pattern of the self-residual subject more deeply, finding the self-residual behavior more efficiently and timely, and has the advantages of practical application.
Description
Technical Field
The invention belongs to the field of social media data mining, and particularly relates to a social media self-mutilation behavior detection method based on strong robustness characteristic selection.
Background
In recent years, self-disabled behavior has gradually become a major challenge in the social public health field. The real need of meeting the challenge is to find the self-disabled behavior in the society timely and effectively. Because the existing traditional self-disability discovery strategy based on the self-disability main body and family friends has the defects of difficult execution, low efficiency and the like, a new self-disability detection strategy is urgently needed. With the popularity of networked social media, more and more people tend to post ideas and record life on social media, thus making self-disabling behavior detection possible using social media. Compared with the traditional self-disabling detection method, the method for detecting the self-disabling behaviors by utilizing the social media can discover more self-disabling behaviors more efficiently.
There is now a great deal of work on social media-based network data to study the health of network users. The prior art provides a psychological pressure detection method based on a heart rate and social media microblog to discover a pressure interval and a pressure source event of a target individual, and the method mainly comprises the following steps: firstly, detecting the abnormal heart rate of an individual to reflect the stress degree of the nervous system of the individual in a test period; then, detecting abnormal intervals of the individual microblogs to find abnormal conditions of frequency of positive microblogs issued by users in a test period; and finally, matching the abnormal heart rate with the abnormal microblog release to determine a pressure interval and find out the pressure source event through microblog data.
The prior art provides a method for early warning of psychological crisis of a social media user, which mainly comprises the following steps: firstly, text data published by a user on a social media is obtained, and the data is preprocessed to obtain a data set formed by words; then, carrying out quantitative emotion analysis calculation on the text through word frequency statistics of the negative words to obtain an emotion feature vector of the text issued by the user; and finally, inputting the obtained feature vector into a neural network to obtain the negative emotion intensity of the user, and grading the psychological state of the user.
The data analysis method based on the social media only selects and uses the homogeneous information source, and does not fully utilize rich heterogeneous information sources on the social media to carry out comprehensive data mining. Meanwhile, the data mining algorithm of the method is too simple, and valuable information in the media data cannot be fully mined and the method is suitable for complex data full of noise in practical application.
Disclosure of Invention
The invention aims to provide a social media self-mutilation behavior detection method based on strong robustness characteristic selection, so as to solve the problems.
In order to achieve the purpose, the invention adopts the following technical scheme:
a social media self-mutilation behavior detection method based on strong robustness feature selection comprises the following steps:
step 1, social media data acquisition: taking historical data of a network social media website as a data source, and acquiring text information, user behavior information, time information and picture information of self-disabled related posts and non-self-disabled posts to obtain a post set consisting of a plurality of posts; post collection composed of n posts
step 3, establishing a self-residual detection model: extracting training samples from the data set constructed in the step 2, and constructing and training a supervised self-residual detection model based on a target function selected by the strong robustness characteristics;
step 4, self-disabled content detection: and (3) constructing a feature vector f of the target post p to be detected according to the feature extraction method in the step (2), inputting the feature vector f into the detection model trained in the step (3) for feature selection, and judging whether the target post p is a self-disabled related post.
Further, in the step 1, in the social media data acquisition, subject crawling of self-disabled related and non-self-disabled posts is performed by using an application program interface provided by a web crawler or a social media through tag information of different social media posts, and main contents acquired for each post include:
(1) Text information: acquiring a title, a subject label word list, a text and all comment texts contained in the text;
(2) User behavior information: acquiring the total posting amount of a posting user, the time for the user to join the social media platform, and the attention number and the fan number of the user;
(3) Time information: acquiring the publishing time of the post and the shooting time of the picture in the post;
(4) Picture information: and acquiring all pictures attached in the posts.
Further, in the step 2 of feature extraction and data set construction, the method mainly comprises the following steps:
(1) Text characteristics: text part of speech distribution characteristics, calculating the proportion of different parts of speech in the text content of each post; readability characteristics, namely calculating readability indexes of the text by using a readability calculation formula in linguistics; the emotion tendency characteristics judge whether the emotion tendency of the post is positive, neutral or negative by utilizing text emotion analysis; word vector representation of the text, calculating the vector representation of each post text by using a depth model; the above feature is represented by w = { w = { [ w ] ling ,w read ,w sent ,w vec }；
(2) The user behavior characteristics are as follows: calculating the average posting volume of the user according to the total posting volume of the user and the time for using the social platform; calculating the average reply rate of the user posts by using the total number of the posts of the user and the number of the posts with replies; plus the user's attention and fan number, it can be characterized as u = { u = } post ,u rep ,u fol ,u fan }；
(3) Time characteristics: each day is divided hourly into 24 time periods, counting the time period of the post publishing time and the shooting time of the attached picture, its characteristics can be expressed as t = { t = } post ,t pic }；
(4) Picture characteristics: representing the color mode in the picture, and simultaneously carrying out quantitative analysis on the emotion dimensionality of the picture by utilizing color information; extracting local features of the picture according to an algorithm in image processing and representing the picture by using a neural network, wherein the features can be expressed as p = { p = { (p) } col ,p sent ,p local ,p net }。
Furthermore, in the step 3 of establishing the self-residual detection model, the high-efficiency and high-robustness self-residual detection model is usedRobust feature selection method: first, use
To represent the annotation information available in the training data, wherein
Chinese post p i When { Y } i1 ＝1,Y i2 =0} the post is a self-disabling content post, otherwise, when { Y } i1 ＝0,Y i2 =1} the post is a normal post;
then, use
A data matrix representing training data, wherein i The number of features extracted for the ith heterogeneous information source;
finally, by using l 2,1 The loss function and the regularization term of the norm achieve the purpose of selecting the strong robustness characteristic; the constructed supervision model is used for training a coefficient matrix
Mapping the data matrix X to a labeling information matrix Y, wherein the training mode is as follows:
wherein,
for the parameters of the regularization term, the specific training process is:
(1) Constructing matrices
Wherein,
is a matrix of the units,
at the same time, the matrix is initialized
Setting a termination threshold value of the convergence of the training process as an element matrix;
(2) Calculation of U = D -1 A T (AD -1 A T ) -1 Y；
(3) Updating diagonal matrix D with diagonal elements of D ii ＝1/(2‖u i ‖ 2 ) Wherein u is i Is Uth row;
(4) Configuration W = (u) 1 ,u 2 ,…,u m-n ) Judging whether the descending amplitude of the target function is less than the epsilon or not, if not, returning to the process (2) to continue training; otherwise, quitting the training and saving the coefficient matrix W.
Further, in the step 4 of self-disabled content detection, a target post p to be detected can be mapped through a learned coefficient matrix W to obtain an indication vector y ∈ R ^2, and when y _1>, y \ u 2, the post is judged to be self-disabled content; otherwise, judging that the post is normal content, wherein the normal content is non-self-disabled content.
Compared with the prior art, the invention has the following technical effects:
(1) Because the reliable help seeking objects are difficult to find in the physical world due to the trust problem, the self-disabled subject is more prone to putting the help seeking to the social media with relative anonymity, so that the self-disabled subject can be more widely contacted by the invention;
(2) The traditional self-disabling behavior research has the defects of few samples, long tracking and observation period and the like, and social media are greatly popularized, and a large amount of accumulated social data contain a large amount of self-disabling cases, so that the self-disabling behavior model can be further explored and understood;
(3) Because of the concealment of the self-disabled behavior, the traditional discovery mode based on relatives and friends of the self-disabled subject is difficult and lagged, and the self-disabled detection mathematical model constructed based on social media data enables the self-disabled behavior to be discovered more timely and effectively.
Drawings
FIG. 1 is a block diagram of a social media self-disabling behavior detection method based on strong robustness feature selection according to the invention.
Fig. 2 is a flow chart of a data acquisition process.
FIG. 3 is a flow chart of a feature analysis process.
FIG. 4 is a flow diagram of a detection model training process.
Fig. 5 is a flow chart of a self-mutilation detection process.
Fig. 6 is an example of self-disabling related posts on social media.
Detailed Description
Embodiments of the present invention will be described in detail below with reference to the drawings and examples. It should be noted that the embodiments described herein are only for explaining the present invention, and are not intended to limit the present invention. Furthermore, the technical features related to the embodiments of the present invention may be combined with each other without conflict.
The specific implementation process of the invention comprises a data acquisition process, a characteristic analysis process, a model establishment process and a self-residual detection process. FIG. 1 is a block diagram of a social media self-disabling behavior detection method based on strong robustness feature selection according to the invention.
1. Data acquisition process
FIG. 6 is an example of social media data for a network. The specific process of data acquisition is as follows:
(1) And (4) crawling the subject according to the label of each data post of the social media through a crawler technology. When crawling from residual relevant content, crawling may be performed using self-residual relevant tags such as "selfherm", "selfinjury", "suicide", and the like; when normal posts are crawled, the target webpage can be crawled in an excessive-demand saturated mode without subjects;
(2) And for the self-residual related posts, considering that the post sets returned by different labels possibly have overlapped parts, carrying out deduplication processing on the self-residual related posts. Then, in order to prevent the situation that non-self-disabled users accidentally publish self-disabled related contents, users and related posts of which the number of posts in a user set to which the self-disabled posts belong is less than 5 are removed;
(3) And filtering the normal content posts through the tags to remove posts with self-disabling related tags. And then, carrying out random sampling according to the demand of normal posts.
The flow of the above steps is shown in FIG. 2, so as to obtain the post set
The crawled content information is shown in fig. 6 and includes:
(1) Text information: acquiring a title, a subject label word list, a text and all comment texts contained in the text;
(2) User behavior information: acquiring the total posting amount of a posting user, the time for the user to join the social media platform, and the attention number and the fan number of the user;
(3) Time information: acquiring the publishing time of the post and the shooting time of the picture in the post;
(4) Picture information: all pictures attached in the post are acquired.
2. Feature analysis process
The posts obtained from the data collection process are subjected to feature analysis and extraction. From post p i (i =1,2, \8230;, n) extracting the features of 4 heterogeneous information sources to obtain a post feature vector f i ＝{w i ,u i ,t i ,p i }. The main process comprises the following steps:
(1) Text feature extraction: the text part-of-speech distribution characteristics are used for calculating the proportion of nouns, verbs, adjectives and adverbs in the text content of each post, and the proportions can be calculated by using a social media-oriented text analysis tool such as CMUTweetTagger; readability characteristics, namely calculating readability indexes of texts by using readability calculation formulas in linguistics, such as a Flesch readability calculation formula, a Linsea Write readability calculation formula, a Fog readability calculation formula and a Dale-Chall readability calculation formula; the emotion tendency characteristics are used for judging whether the emotion tendency of the post is positive, neutral or negative by utilizing text emotion analysis and can be calculated by using corpus MPQA; word vector of textRepresenting, a vector representation of each post is computed for its text using the word2vec model. The above feature is denoted by w i ＝{w ling ,w read ,w sent ,w vec }；
(2) Extracting user behavior characteristics: calculating the average posting volume of the user according to the total posting volume of the user and the time for using the social platform; calculating the average reply rate of the user posts by using the total number of the posts of the user and the number of the posts with replies; plus the user's attention count and fan count, which may be characterized as u i ＝{u post ,u rep ,u fol ,u fan }；
(3) Time characteristic extraction: each day is divided into 24 time periods by hour, and the time period of the post issuance time and the time period of the attached picture taking time are counted, so that {0,1} can be used 24 Is characterized by a vector of (a), whose features can be represented as t i ＝{t post ,t pic }；
(4) Extracting picture features: representing the color pattern in the picture by using a cylindrical coordinate color space HSV to obtain the chroma (Hue), the color Saturation (Saturation) and the Brightness (Brightness) of the picture. Meanwhile, quantitative analysis is carried out on the emotion dimensionality of the picture by utilizing the color information, and the calculation formula is as follows:
in addition, local features of the pictures are extracted using SURF, LBP and GIST algorithms in image processing, and feature extraction is performed on the pictures using AlexNet neural networks that have been pre-trained on ImageNet datasets. Which can be characterized by p i ＝{p col ,p sent ,p local ,p net }。
A flow chart of this process is shown in fig. 3. Thereafter, a self-incomplete post data set and a normal post data set may be constructed, respectively.
3. Model building Process
Defining self-disabling content annotations in training datasetThe information is
Wherein, it is to
Chinese post p i When { Y } i1 ＝1,Y i2 =0} the post is a self-disabled content post; otherwise, when { Y } i1 ＝0,Y i2 =1} the post is a normal post. Data matrix composed using feature vectors of training data
(l i The number of features extracted for the ith heterogeneous information source), the constructed supervised model based on the selection of the strong robustness features is used for training a coefficient matrix
Mapping the data matrix X to a labeling information matrix Y, wherein the training mode is as follows:
wherein,
for regularization term parameters, the specific training process is:
(1) Constructing matrices
Wherein,
is a matrix of the unit, and is,
at the same time, the matrix is initialized
Is a unit matrix, and sets the termination condition of the convergence of the training as (II XW-Y II) 2,1 +αW2,1<∈；
(2) Calculate U = D -1 A T (AD -1 A T ) -1 Y；
(3) Updating diagonal matrix D with diagonal elements of D ii ＝1/(2‖u i ‖ 2 ) Wherein u is i Is Uth line;
(4) Configuration W = (u) 1 ,u 2 ,…,u m-n ) Judging whether the termination condition is satisfied, if not, returning to the process (2) to continue training; otherwise, exiting from training and saving the coefficient matrix W.
The flow chart of the training process of the supervised model is shown in fig. 4.
4. Self-mutilation detection process
And (3) for the target post p needing to be detected, constructing a feature vector f according to a feature extraction method in feature analysis, inputting the feature vector f into a detection model obtained by training in the model establishing process, and judging whether the target post p is a self-residual related post. By mapping the coefficient matrix W of the supervised self-residual detection model, an indication vector of the target post p can be obtained
When y is 1 >y 2 If so, judging the post as the self-disabled content; otherwise, judging the post as normal content. The detection process flow diagram is shown in fig. 5.
Claims (1)
1. A social media self-disabling behavior detection method based on strong robustness feature selection is characterized by comprising the following steps:
step 1, social media data acquisition: taking historical data of a network social media website as a data source, and acquiring text information, user behavior information, time information and picture information of self-disabled related posts and non-self-disabled posts to obtain a post set consisting of a plurality of posts; post set composed of n posts
Step 2, data feature extraction and data set construction: for posts p obtained from data collection i Wherein i =1,2, ·, n; extracting the characteristics of 4 heterogeneous information sources to obtain a post characteristic vector f i ＝{w i ，u i ，t i ，p i In which w i Representing a text feature, u i Representing a user behavior feature, t i Representing the temporal characteristics of the post, p i Representing picture characteristics of the posts, and respectively constructing a self-disabled post data set and a normal post data set;
step 3, establishing a self-residual detection model: extracting training samples from the data set constructed in the step 2, and constructing and training a supervised self-residual detection model based on a target function selected by the strong robustness characteristics;
step 4, self-disabled content detection: constructing a feature vector f of a target post p to be detected according to the feature extraction method in the step 2, inputting the feature vector f into the detection model trained in the step 3 for feature selection, and judging whether the target post p is a self-mutilation related post;
in step 1, in social media data acquisition, through tag information of different social media posts, topic crawling of self-residual related posts and non-self-residual posts is performed by using an application program interface provided by a web crawler or a social media, and main contents acquired by each post comprise:
(1) Text information: acquiring a title, a subject label word list, a text and all comment texts contained in the text;
(2) User behavior information: acquiring the total posting volume of posting users, the time for the users to join the social media platform, and the attention number and the fan number of the users;
(3) Time information: acquiring the publishing time of the post and the shooting time of the picture in the post;
(4) Picture information: acquiring all pictures attached to the posts;
step 2, in feature extraction and data set construction, the method mainly comprises the following steps:
(1) Text characteristics: text part-of-speech distribution feature w ling Calculating the proportion of different parts of speech in the text content of each post; readability characteristic w read Calculating the readability index of the text by using a readability calculation formula in linguistics; emotional tendency characteristics w sent Judging whether the emotional tendency of the post is positive, neutral or negative by utilizing text emotional analysis; word vector representation w of text vec Calculating vector representation of the text of each post by using a depth model; the above feature is represented by w = { w = { [ w ] ling ，w read ，w sent ，w vec }；
(2) The user behavior characteristics are as follows: calculating the average posting volume u of the user according to the total posting volume of the user and the time of using the network social media post (ii) a Calculating the average reply rate u of the user posts by using the total number of the posts of the user and the number of the posts with replies rep (ii) a Plus the user's attention number u fol Number u of vermicelli made from bean starch fan Its characteristic can be expressed as u = { u = { u = post ，u rep ，u fol ，u fan }；
(3) Time characteristics: dividing each day into 24 time periods by hour, and counting the post publishing time t post And a time period t of the shooting time of the attached picture pic Characterized in that it can be represented as t = { t = { t = } post ，t pic }；
(4) Picture characteristics: for color pattern p in picture col Representing and utilizing color information to carry out emotion dimension p on the picture sent Carrying out quantitative analysis; local characteristic p of picture according to algorithm in image processing local Extracting and representing the picture p by using a neural network net Its characteristics can be expressed as p = { p = col ，p sent ，p local ，p net }；
Step 3, in the self-residual detection model establishment, a high-efficiency and steady feature selection method with strong robustness is used: first, use
To represent the annotation information available in the training data, wherein
Chinese post p i When { Y } i1 ＝1，Y i2 =0} the post is a self-spoiled content post, otherwise, when { Y } i1 ＝0，Y i2 =1}, the post is a normal post;
then, use
A data matrix representing training data, wherein i The feature quantity extracted for the ith heterogeneous information source;
finally, by using l 2，1 The loss function and the regularization term of the norm achieve the purpose of selecting the strong robustness characteristics; the constructed supervision model is used for training a coefficient matrix
Mapping the data matrix X to a labeling information matrix Y, wherein the training mode is as follows:
wherein,
for the parameters of the regularization term, the specific training process is:
(1) Constructing matrices
Wherein,
is a matrix of the units,
at the same time, the matrix is initialized
Setting a termination threshold value of the convergence of the training process as an element matrix;
(2) Computing
(3) Updating diagonal matrix D with diagonal elements D ii ＝1/(2||u i || 2 ) Wherein u is i Is Uth line;
(4) Configuration W = (u) 1 ，u 2 ，...，u m-n ) Judging whether the descending amplitude of the target function is less than the epsilon or not, if not, returning to the process (2) to continue training; otherwise, quitting training and saving the coefficient matrix W;
step 4, in the self-disabled content detection, a target post p to be detected can be mapped through a learned coefficient matrix W to obtain an indication vector y ∈ R ^2, and when y _1 is larger than y _2, the post is judged to be self-disabled content; otherwise, judging that the post is normal content, wherein the normal content is non-self-disabled content.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911033392.0A CN110867250B (en) | 2019-10-28 | 2019-10-28 | Social media self-disabling behavior detection method based on strong robustness feature selection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911033392.0A CN110867250B (en) | 2019-10-28 | 2019-10-28 | Social media self-disabling behavior detection method based on strong robustness feature selection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110867250A CN110867250A (en) | 2020-03-06 |
| CN110867250B true CN110867250B (en) | 2022-10-25 |
Family
ID=69653491
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911033392.0A Active CN110867250B (en) | 2019-10-28 | 2019-10-28 | Social media self-disabling behavior detection method based on strong robustness feature selection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110867250B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107145524A (en) * | 2017-04-12 | 2017-09-08 | 清华大学 | Suicide risk checking method and system based on microblogging and Fuzzy Cognitive Map |
| CN109903851A (en) * | 2019-01-24 | 2019-06-18 | 暨南大学 | A kind of automatic Observation technology of the psychological abnormality variation based on social networks |
| CN110263620A (en) * | 2019-05-06 | 2019-09-20 | 杭州电子科技大学 | One kind being based on L2,1The age estimation method of label study partially |
-
2019
- 2019-10-28 CN CN201911033392.0A patent/CN110867250B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107145524A (en) * | 2017-04-12 | 2017-09-08 | 清华大学 | Suicide risk checking method and system based on microblogging and Fuzzy Cognitive Map |
| CN109903851A (en) * | 2019-01-24 | 2019-06-18 | 暨南大学 | A kind of automatic Observation technology of the psychological abnormality variation based on social networks |
| CN110263620A (en) * | 2019-05-06 | 2019-09-20 | 杭州电子科技大学 | One kind being based on L2,1The age estimation method of label study partially |
Non-Patent Citations (2)
| Title |
|---|
| Latent Suicide Risk Detection on Microblog via Suicide-OrientedWord Embeddings and Layered Attention;Lei Cao 等;《arXiv:1910.12038v1》;20191026;第1-11页 * |
| 基于中文微博语言特征的自杀意念检测;许立鹏 等;《中北大学学报（自然科学版）》;20190611;第40卷(第4期);第350-357页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110867250A (en) | 2020-03-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109783614B (en) | Differential privacy disclosure detection method and system for to-be-published text of social network | |
| CN104573130B (en) | The entity resolution method and device calculated based on colony | |
| CN108549647B (en) | Method for realizing active prediction of emergency in mobile customer service field without marking corpus based on SinglePass algorithm | |
| Perdana et al. | Combining likes-retweet analysis and naive bayes classifier within twitter for sentiment analysis | |
| CN112199608A (en) | Social media rumor detection method based on network information propagation graph modeling | |
| Guthier et al. | Detection and visualization of emotions in an affect-aware city | |
| CN112559684A (en) | Keyword extraction and information retrieval method | |
| CN111159763A (en) | System and method for analyzing portrait of law-related personnel group | |
| US20200004738A1 (en) | Generating further knowledge to process query | |
| Han et al. | Linking social network accounts by modeling user spatiotemporal habits | |
| CN113422761A (en) | Malicious social user detection method based on counterstudy | |
| CN109086794A (en) | A kind of driving behavior mode knowledge method based on T-LDA topic model | |
| CN109949174A (en) | A kind of isomery social network user entity anchor chain connects recognition methods | |
| CN109271546A (en) | The foundation of image retrieval Feature Selection Model, Database and search method | |
| CN114329455B (en) | User abnormal behavior detection method and device based on heterogeneous graph embedding | |
| CN108153818B (en) | Big data based clustering method | |
| CN116910628B (en) | Creator expertise portrait assessment method and system | |
| CN113158672A (en) | Relation analysis method and device based on news events | |
| CN106844743B (en) | Emotion classification method and device for Uygur language text | |
| CN110867250B (en) | Social media self-disabling behavior detection method based on strong robustness feature selection | |
| CN108647739A (en) | A kind of myspace discovery method based on improved density peaks cluster | |
| WO2021081741A1 (en) | Image classification method and system employing multi-relationship social network | |
| TWI761090B (en) | Dialogue data processing system and method thereof and computer readable medium | |
| CN112445895B (en) | Method and system for identifying user search scene | |
| ElGindy et al. | Enriching user profiles using geo-social place semantics in geo-folksonomies |
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 |