A system and method for character label profiling based on network behavior traces

Abstract

The present application relates to the technical field of behavior analysis, in particular to a system and method for character label characterization based on network behavior traces, the characterization method comprising the following steps: obtaining network behavior traces when a user historically browses a page, capturing keywords in the historically browsed page, and comparing the keywords with keywords in an established database; screening invalid behavior traces according to the historically browsed page, and excluding invalid behavior traces to obtain valid behavior traces; setting a plurality of character labels containing corresponding keywords, dividing the valid behavior traces according to the character labels, taking a character label corresponding to the valid behavior trace with the highest proportion as a character label of the user, and evaluating a risk value of the user under the character label; restricting the network behavior of the user according to the risk level of the user; and adjusting the character label of the user according to the total number of valid behavior traces corresponding to different character labels.

Description

Technical Field

[0001] This invention relates to the field of behavioral analysis technology, specifically to a system and method for characterizing individuals based on online behavioral traces. Background Technology

[0002] In today's era, the internet has permeated all aspects of social life, greatly enriching people's daily lives. However, at the same time, inappropriate dating apps, online loans, online gambling, and live streaming apps have also surged with the mobile internet wave, seriously affecting people's lives. Therefore, it is necessary to monitor users' online behavior to prevent them from falling into these traps and suffering harm.

[0003] Conventional online behavior monitoring captures all user actions, but because users may accidentally enter certain pages, the monitored online behavior may be inaccurate, and the task tags assigned to users may also be inaccurate. At the same time, users' online behavior traces are affected by multiple factors, and the generated user tags may no longer accurately reflect users' online behavior after a certain period of time. Summary of the Invention

[0004] The purpose of this invention is to provide a system and method for characterizing individuals based on online behavior traces, in order to solve the problems mentioned in the background art.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: a method for characterizing online behavior traces, the method comprising the following steps:

[0006] Step S100: Obtain the network behavior traces of the user's historical browsing pages, capture the keywords in the historical browsing pages, compare the keywords with the keywords in the established database to obtain normal pages and abnormal pages; filter the invalid behavior traces in the network behavior traces based on the historical browsing pages, and eliminate invalid behavior traces to obtain valid behavior traces;

[0007] Step S200: Set several person tags containing corresponding keywords, divide the effective behavioral traces according to the set person tags, take the person tag corresponding to the most effective behavioral traces as the user's person tag, and assess the user's risk value under the person tag.

[0008] Step S300: Based on the comparison between the user's risk level and the set danger threshold, determine whether to designate the person tag and user as a target of attention and restrict the user's online behavior;

[0009] Step S400: Monitor the user's online behavior traces in real time within a period, and adjust the user's persona tags according to the total number of valid behavior traces corresponding to different persona tags.

[0010] Furthermore, step S100 includes the following steps:

[0011] Step S101: Establish two databases, one database containing several keywords corresponding to the abnormal page, and the other database containing several keywords corresponding to the normal page;

[0012] Step S102: Read the browsing duration and number of times the user has viewed historical pages. Let the i-th network behavior trace be A. i Network Behavior Traces A i The browsing time for the corresponding page is t. i and the number of views is n i Set a threshold ε for the duration of a single browsing session; when t i When the value is greater than ε, then the network behavior trace A will be recorded. i As a valid record of behavior;

[0013] Step S103: When t i When ≤ε, read the (i-1)th network behavior trace A. i-1 From network behavior traces A i-1 Keywords are captured from the corresponding browsing pages, and these keywords are compared with keywords in two databases to determine online behavior traces A. i-1 Is the corresponding page an abnormal page? According to the formula:

[0014]

[0015] Where IF() is the judgment function, and N is the total number of historical network behavior traces; when network behavior trace A i-1 The corresponding page to be viewed is an abnormal page, IF(A i-1 When network behavior trace A = 1, i-1 When the corresponding page being viewed is a normal page, IF(A i-1 ) = 0;

[0016] Network behavior traces A i-1 The purpose of identifying abnormal pages is to prevent users from accidentally entering abnormal pages and having their network behavior traces marked as valid, thus affecting subsequent filtering results.

[0017] Calculate network behavior trace A i The risk value D; set a risk threshold ∈, when D≥∈, then the network behavior trace A i As a valid behavioral trace, when D < ∈, then the network behavioral trace A is considered valid. i As a record of invalid behavior.

[0018] Furthermore, step S200 includes the following steps:

[0019] Step S201: Set p individual character tags, where each character tag contains several words corresponding to the character tag, capture the words in the browsing page corresponding to the effective behavior traces and compare them with the words in the p individual character tags, and use the character tags as the basis for classifying the effective behavior traces;

[0020] Step S202: Capture vocabulary for the browsing page corresponding to each valid behavioral trace. Let the j-th valid behavioral trace be X. j The qth character is tagged as B. q When from valid behavioral traces X j The corresponding person tag B was captured from the browsing page. q Then the character tag B q Marked to valid behavioral trace X j In the middle; when from the effective behavioral trace X j The words captured on the corresponding browsing page correspond to the character tag B in the p character tags. q And character tag B o B o If the o-th character tag is among p character tags, then continue with the (j+1)-th valid behavioral trace X. j+1 Vocabulary capture is performed on the corresponding browsing page; when effective behavioral traces X j+1 The corresponding person tag B was captured from the browsing page. q Then the character tag B q Marked to valid behavioral trace X j And Effective Behavioral Traces X j+1 middle;

[0021] Step S203: Count the number of entries for each person tag and select the person tag with the most valid behavioral traces as the user's person tag;

[0022] Step S204: Let the user's character tag be B. q Tag B for the person q The number of valid behavioral traces is N q Set N q The bar is marked as B q The uth valid behavior trace in the valid behavior traces is (X) q ) u Valid behavioral traces (X) q ) u The browsing time is and number of views According to the formula:

[0023]

[0024] Where N′ is the total number of valid behavioral traces, t j For the j-th valid behavioral trace X j Browsing duration, n j For the j-th valid behavioral trace X j The number of views; calculate the person tag B q proportion of all valid behavioral traces

[0025] Depending on the user's persona, the weight of each persona in the user's risk assessment varies; adjusting the weight of each persona can more accurately assess the user's risk level.

[0026] Step S205: In the p character tags, remove the q-th character tag B. q The remaining p-1 person tags Among them, B α Let α be the tag of the αth person and α ≠ p; according to the formula:

[0027]

[0028] Where v represents the v-th valid behavioral trace in any character tag, and N α Let α be the number of valid behavioral traces under the αth person tag; calculate the number of traces under person tag B. q The user's risk value F q .

[0029] Furthermore, step S300 includes the following steps:

[0030] Step S301: Monitor the user's online behavior traces and capture the vocabulary of the pages browsed corresponding to the user's online behavior traces;

[0031] Step S302: Set a danger threshold δ and set the character tag B. q As a user's persona, and in persona B q The risk value of the user is F q When F q When the value is ≥δ, then the user and person tags B will be applied. q Listed as a target of attention, if any word corresponding to a person's tag is detected, the user's online behavior will be restricted; when F q When <δ, the user's online behavior traces are captured, corresponding to words on the browsed page and the person tag B. qIf the user is not detected, a warning will be issued. The duration t′ of the network behavior trace will be recorded, and a time limit τ will be set. When t′≥τ, the user's network behavior will be restricted.

[0032] After identifying a user's profile, we will continue to monitor their online behavior and impose restrictions on them if they continue such online activities.

[0033] Furthermore, step S400 includes the following steps:

[0034] Step S401: Monitor the user's online behavior traces in real time, assuming that the q-th person tag B is captured in one unit period T. q The number of valid behavioral traces is n q Then, after one period T, the qth person's tag B q Total number of valid behavioral traces (n) q ) total =N q +n q In the p individual character tags, besides the character tag B q The αth character tag B, besides α The number of valid behavioral traces is n α Then, after one cycle T, the character tag B α Total number of valid behavioral traces (n) α ) total =N α +n α ;

[0035] Step S402: When (n q ) total ≥(n α ) total If so, continue to use character tag B. q As a character's tag; when α = α1 Then tag the person As a character's label;

[0036] By monitoring users' online behavior in real time, we can dynamically adjust their profile tags to ensure that they are always accurately portrayed.

[0037] Furthermore, in order to better implement the above method, a character tagging system based on online behavior traces is proposed. The characterization system includes an online behavior trace screening module, a risk assessment module, an online behavior warning module, and a character tag adjustment module.

[0038] The network behavior trace filtering module is used to obtain the network behavior traces of users when browsing historical pages, capture keywords in the historical browsing pages, compare the keywords with the keywords in the established database to obtain normal pages and abnormal pages; and filter invalid behavior traces in the network behavior traces based on the historical browsing pages to obtain valid behavior traces.

[0039] The risk assessment module is used to set several person tags containing corresponding keywords, divide the effective behavioral traces according to the set person tags, take the person tag corresponding to the most effective behavioral traces as the user's person tag, and assess the user's risk value under the person tag.

[0040] The network behavior warning module is used to determine whether to designate the person tag and user as a target of attention based on the comparison result between the user's risk level and the set danger threshold, and to restrict the user's network behavior.

[0041] The user tag adjustment module is used to monitor a user's online behavior traces in real time over a period of time, and adjust the user's user tags according to the total number of valid behavior traces corresponding to different user tags.

[0042] Furthermore, the network behavior trace filtering module includes a network behavior trace segmentation unit and a target behavior trace filtering unit;

[0043] The network behavior trace segmentation unit is used to acquire network behavior traces of users browsing historical pages, capture keywords in historical browsing pages, compare the keywords with keywords in the established database, and obtain normal pages and abnormal pages; the target behavior trace filtering unit is used to filter invalid behavior traces in network behavior traces based on historical browsing pages, and exclude invalid behavior traces to obtain valid behavior traces.

[0044] Furthermore, the risk assessment module includes a person tag acquisition unit and a risk calculation unit;

[0045] The user tag acquisition unit is used to set several user tags containing corresponding keywords, divide the valid behavioral traces according to the set user tags, and take the user tag corresponding to the most frequent valid behavioral traces as the user's user tag; the risk calculation unit is used to assess the user's risk value under the user tags.

[0046] Compared with the prior art, the beneficial effects achieved by the present invention are: (1) Before continuing to characterize the user's network behavior traces, the present invention screens the network behavior traces and excludes some network behavior traces that are irrelevant to the characterization of character tags, thereby improving work efficiency and making the definition of the user's character tags more accurate; (2) After determining the user's character tags, the present invention dynamically adjusts the weight of all character tags under different character tags in risk assessment, so that the character tags under the user's character tags can have a greater impact on the risk assessment results, and strengthens the importance attached to the user's character tags; (3) After determining the user's character tags, the present invention will continue to monitor the user's network behavior traces and constrain the user when the user continues related network behaviors; (4) The present invention monitors the user's network behavior traces in real time and can dynamically adjust the user's character tags to ensure that the user's character tags are always correctly characterized. Attached Figure Description

[0047] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0048] Figure 1 This is a step-by-step diagram of a method for characterizing individuals based on online behavioral traces;

[0049] Figure 2 This is a schematic diagram of a character tagging system based on online behavior traces. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] Please see Figure 1 , Figure 2 This invention provides a technical solution: a method for characterizing online behavior traces, the method comprising the following steps:

[0052] Step S100: Obtain the network behavior traces of the user's historical browsing pages, capture the keywords in the historical browsing pages, compare the keywords with the keywords in the established database to obtain normal pages and abnormal pages; filter the invalid behavior traces in the network behavior traces based on the historical browsing pages, and eliminate invalid behavior traces to obtain valid behavior traces;

[0053] Step S100 includes the following steps:

[0054] Step S101: Establish two databases, one database containing several keywords corresponding to the abnormal page, and the other database containing several keywords corresponding to the normal page;

[0055] Step S102: Read the browsing duration and number of times the user has viewed the historical pages. Let the i-th network behavior trace be A. i Network Behavior Traces A i The browsing time for the corresponding page is t. i and the number of views is n i Set a threshold ε for the duration of a single browsing session. When t i When the value is greater than ε, then the network behavior trace A will be recorded. i As a valid record of behavior;

[0056] Step S103: When t i When ≤ε, read the (i-1)th network behavior trace A. i-1 From network behavior traces A i-1 Keywords are captured from the corresponding browsing pages, and these keywords are compared with keywords in two databases to determine online behavior traces A. i-1 Is the corresponding page an abnormal page? According to the formula:

[0057]

[0058] Where IF() is the conditional function, and N is the total number of historical network behavior traces; when network behavior trace A i-1 The corresponding page to be viewed is an abnormal page, IF(A i-1 When network behavior trace A = 1, i-1 When the corresponding page being viewed is a normal page, IF(A i-1 ) = 0;

[0059] Calculate network behavior trace A i The risk value D; set a risk threshold ∈, when D≥∈, then the network behavior trace A i As a valid behavioral trace, when D < ∈, then the network behavioral trace A is considered valid. i As a record of invalid behavior.

[0060] Step S200: Set several person tags containing corresponding keywords, divide the effective behavioral traces according to the set person tags, take the person tag corresponding to the most effective behavioral traces as the user's person tag, and assess the user's risk value under the person tag.

[0061] Step S200 includes the following steps:

[0062] Step S201: Set p individual character tags, where each character tag contains several words corresponding to the character tag, capture the words in the browsing page corresponding to the effective behavior traces and compare them with the words in the p individual character tags, and use the character tags as the basis for classifying the effective behavior traces;

[0063] Step S202: Capture vocabulary for the browsing page corresponding to each valid behavioral trace. Let the j-th valid behavioral trace be X. j The qth character is tagged as B. q When from valid behavioral traces X j The corresponding person tag B was captured from the browsing page. q Then the character tag B q Marked to valid behavioral trace X j In the middle; when from the effective behavioral trace X j The words captured on the corresponding browsing page correspond to the character tag B in the p character tags. q And character tag B o B o If the o-th character tag is among p character tags, then continue with the (j+1)-th valid behavioral trace X. j+1 Vocabulary capture is performed on the corresponding browsing page; when effective behavioral traces X j+1 The corresponding person tag B was captured from the browsing page. q Then the character tag B q Marked to valid behavioral trace X j And Effective Behavioral Traces X j+1 middle;

[0064] Step S203: Count the number of entries for each person tag and select the person tag with the most valid behavioral traces as the user's person tag;

[0065] Step S204: Let the user's character tag be B. q Tag B for the person q The number of valid behavioral traces is N q Set N q The bar is marked as B q The uth valid behavior trace in the valid behavior traces is (X) q ) u Valid behavioral traces (X) q ) u The browsing time is and number of views According to the formula:

[0066]

[0067] Where N′ is the total number of valid behavioral traces, t j For the j-th valid behavioral trace X j Browsing duration, n j For the j-th valid behavioral trace X j The number of views; calculate the person tag B q proportion of all valid behavioral traces

[0068] Step S205: In the p character tags, remove the q-th character tag B. q The remaining p-1 person tags Among them, B α Let α be the tag of the αth person and α ≠ p; according to the formula:

[0069]

[0070] Where v represents the v-th valid behavioral trace in any character tag, and N α Let α be the number of valid behavioral traces under the αth person tag; calculate the number of traces under person tag B. q The user's risk value F q .

[0071] Step S300: Based on the comparison between the user's risk level and the set danger threshold, determine whether to designate the person tag and user as a target of attention and restrict the user's online behavior;

[0072] Step S300 includes the following steps:

[0073] Step S301: Monitor the user's online behavior traces and capture the vocabulary of the pages browsed corresponding to the user's online behavior traces;

[0074] Step S302: Set a danger threshold δ and set the character tag B. q As a user's persona, and in persona B q The risk value of the user is F q When F q When the value is ≥δ, then the user and person tags B will be applied. q Listed as a target of attention, if any word corresponding to a person's tag is detected, the user's online behavior will be restricted; when F q When <δ, the user's online behavior traces are captured, corresponding to words on the browsed page and the person tag B. q If the user is not detected, a warning will be issued. The duration t' of the network behavior trace will be recorded, and a time limit τ will be set. When t' ≥ τ, the user's network behavior will be restricted.

[0075] Step S400: Monitor the user's online behavior traces in real time within a period, and adjust the user's persona tags according to the total number of valid behavior traces corresponding to different persona tags;

[0076] Step S400 includes the following steps:

[0077] Step S401: Monitor the user's online behavior traces in real time, assuming that the q-th person tag B is captured in one unit period T. q The number of valid behavioral traces is n q Then, after one period T, the qth person's tag B q Total number of valid behavioral traces (n) q ) total =N q +n q In the p individual character tags, besides the character tag B q The αth character tag B, besides α The number of valid behavioral traces is n α Then, after one cycle T, the character tag B α Total number of valid behavioral traces (n) α ) total =N α +n α ;

[0078] Step S402: When (n q ) total ≥(n α ) total If so, continue to use character tag B. q As a character's tag; when α = α1 Then tag the person As a character label.

[0079] A character tagging system based on online behavior traces, the system comprising an online behavior trace screening module, a risk assessment module, an online behavior warning module, and a character tag adjustment module;

[0080] The network behavior trace filtering module is used to obtain the network behavior traces of users when browsing historical pages, capture keywords in the historical browsing pages, compare the keywords with the keywords in the established database to obtain normal pages and abnormal pages; and filter invalid behavior traces in the network behavior traces based on the historical browsing pages to obtain valid behavior traces.

[0081] The risk assessment module is used to set several person tags containing corresponding keywords, divide the effective behavioral traces according to the set person tags, take the person tag corresponding to the most effective behavioral traces as the user's person tag, and assess the user's risk value under the person tag.

[0082] The network behavior warning module is used to determine whether to designate the person tag and user as a target of attention based on the comparison result between the user's risk level and the set danger threshold, and to restrict the user's network behavior.

[0083] The user tag adjustment module is used to monitor a user's online behavior traces in real time over a period of time, and adjust the user's user tags according to the total number of valid behavior traces corresponding to different user tags.

[0084] The network behavior trace filtering module includes a network behavior trace segmentation unit and a target behavior trace filtering unit.

[0085] The network behavior trace segmentation unit is used to acquire network behavior traces of users browsing historical pages, capture keywords in historical browsing pages, compare the keywords with keywords in the established database, and obtain normal pages and abnormal pages; the target behavior trace filtering unit is used to filter invalid behavior traces in network behavior traces based on historical browsing pages, and exclude invalid behavior traces to obtain valid behavior traces.

[0086] The risk assessment module includes a person tag acquisition unit and a risk calculation unit.

[0087] The user tag acquisition unit is used to set several user tags containing corresponding keywords, divide the valid behavioral traces according to the set user tags, and take the user tag corresponding to the most frequent valid behavioral traces as the user's user tag; the risk calculation unit is used to assess the user's risk value under the user tags.

[0088] Two databases were established. One database contained keywords corresponding to abnormal pages, such as "loan," "gambling," "live streaming," and "dating." The other database contained keywords corresponding to normal pages. Historical network behavior traces were read. There were 12 network behavior traces in total. The second network behavior trace had a browsing duration of 1 second and was viewed once. A single browsing duration threshold ε = 5 seconds was set. Since 1 second < 5 seconds, the keywords of the first network behavior trace were read to determine if they contained keywords corresponding to abnormal pages. Because they did not contain keywords corresponding to abnormal pages, IF( A1 ) = 0, The risk threshold is set to ∈ = 5s, therefore the second network behavior trace is an invalid behavior trace; because the first network behavior trace does not contain the negative keywords of the corresponding abnormal page, it is also an invalid behavior trace; thus, 10 valid behavior traces are obtained.

[0089] Four custom user tags were defined: loan, gambling, live streaming, and dating. Ten valid behavioral traces were collected. Among them, six traces involved loans, two involved gambling, one involved live streaming, and one involved dating. Therefore, loan was selected as the user's user tag.

[0090] Assume that each valid behavioral trace involving loans is viewed once, with viewing durations of 5s, 7s, 8s, 10s, 15s, and 20s respectively; each valid behavioral trace involving gambling is viewed once, with viewing durations of 5s and 10s respectively; each valid behavioral trace involving live streaming is viewed once, with a viewing duration of 20s; and each valid behavioral trace involving dating is viewed once, with a viewing duration of 10s.

[0091] The calculated proportion involved loans is: Therefore, the remaining three character tags have the following weight: Therefore, the risk value F is calculated;

[0092]

[0093] A danger threshold δ = 50 is set. Since 44.515 < 50, a warning will be issued when the user continues to engage in loan-related online activities. If the user's browsing time exceeds 10 seconds, the user's online activities will be restricted.

[0094] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0095] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for characterizing individuals based on online behavioral traces, characterized in that: The depiction method includes the following steps: Step S100: Obtain the network behavior traces of the user's historical browsing pages, capture the keywords in the historical browsing pages, compare the keywords with the keywords in the established database to obtain normal pages and abnormal pages; filter the invalid behavior traces in the network behavior traces based on the historical browsing pages, and eliminate invalid behavior traces to obtain valid behavior traces; Step S200: Set several person tags containing corresponding keywords, divide the effective behavioral traces according to the set person tags, take the person tag corresponding to the most effective behavioral traces as the user's person tag, and assess the user's risk value under the person tag. Step S300: Based on the comparison between the user's risk level and the set danger threshold, determine whether to designate the person tag and user as a target of attention and restrict the user's online behavior; Step S400: Monitor the user's online behavior traces in real time within a period, and adjust the user's persona tags according to the total number of valid behavior traces corresponding to different persona tags.

2. The method for characterizing individuals based on online behavior traces according to claim 1, characterized in that: Step S100 includes the following steps: Step S101: Establish two databases, one database containing several keywords corresponding to the abnormal page, and the other database containing several keywords corresponding to the normal page; Step S102: Read the browsing duration and number of times the user has viewed historical pages. Let the i-th network behavior trace be A. i Network Behavior Traces A i The browsing time for the corresponding page is t. i and the number of views is n i Set a threshold ε for the duration of a single browsing session; when t i When the value is greater than ε, then the network behavior trace A will be recorded. i As a valid record of behavior; Step S103: When t i When ≤ε, read the (i-1)th network behavior trace A. i-1 From network behavior traces A i-1 Keywords are captured from the corresponding browsing pages, and these keywords are compared with keywords in two databases to determine online behavior traces A. i-1 Is the corresponding page an abnormal page? According to the formula: Where IF() is the conditional function, and N is the total number of historical network behavior traces; when network behavior trace A i-1 The corresponding page to be viewed is an abnormal page, IF(A i-1 When network behavior trace A = 1, i-1 When the corresponding page being viewed is a normal page, IF(A i-1 ) = 0; Calculate network behavior trace A i The risk value D; set a risk threshold ∈, when D≥∈, then the network behavior trace A i As valid behavioral traces, when D < ∈, network behavioral trace Ai is considered invalid behavioral trace.

3. The method for characterizing individuals based on online behavior traces according to claim 2, characterized in that: Step S200 includes the following steps: Step S201: Set p individual character tags, where each character tag contains several words corresponding to the character tag, capture the words in the browsing page corresponding to the effective behavior traces and compare them with the words in the p individual character tags, and use the character tags as the basis for classifying the effective behavior traces; Step S202: Capture vocabulary for the browsing page corresponding to each valid behavioral trace. Let the j-th valid behavioral trace be X. j The qth character is tagged as B. q When from valid behavioral traces X j The corresponding person tag B was captured from the browsing page. q Then the character tag B q Marked to valid behavioral trace X j In the middle; when from the effective behavioral trace X j The words captured on the corresponding browsing page correspond to the character tag B in the p character tags. q And character tag B o B o If the o-th character tag is among p character tags, then continue with the (j+1)-th valid behavioral trace X. j+1 Vocabulary capture is performed on the corresponding browsing page; when effective behavioral traces X j+1 The corresponding person tag B was captured from the browsing page. q Then the character tag B q Marked to valid behavioral trace X j And Effective Behavioral Traces X j+1 middle; Step S203: Count the number of entries for each person tag and select the person tag with the most valid behavioral traces as the user's person tag; Step S204: Let the user's character tag be B. q Tag B for the person q The number of valid behavioral traces is N q Set N q The bar is marked as B q The uth valid behavior trace in the valid behavior traces is (X) q ) u Valid behavioral traces (X) q ) u The browsing time is and number of views According to the formula: Where N′ is the total number of valid behavioral traces, t j For the j-th valid behavioral trace X j Browsing duration, n j For the j-th valid behavioral trace X j The number of views; calculate the person tag B q proportion of all valid behavioral traces Step S205: In the p character tags, remove the q-th character tag B. q The remaining p-1 person tags Among them, B α Let α be the tag of the αth person and α ≠ p; according to the formula: Where v represents the v-th valid behavioral trace in any character tag, and N α Let α be the number of valid behavioral traces under the αth person tag; calculate the number of traces under person tag B. q The user's risk value F q .

4. The method for characterizing individuals based on online behavior traces according to claim 3, characterized in that: Step S300 includes the following steps: Step S301: Monitor the user's online behavior traces and capture the vocabulary of the pages browsed corresponding to the user's online behavior traces; Step S302: Set a danger threshold δ and set the character tag B. q As a user's persona, and in persona B q The risk value of the user is F q When F q When the value is greater than or equal to δ, then the user and person tags B will be applied. q Listed as a target of attention, if any word corresponding to a person's tag is detected, the user's online behavior will be restricted; when F q When <δ, the user's online behavior traces are captured, corresponding to words on the browsed page and the person tag B. q If the user is not detected, a warning will be issued. The duration t′ of the network behavior trace will be recorded, and a time limit τ will be set. When t′≥τ, the user's network behavior will be restricted.

5. The method for characterizing individuals based on online behavior traces according to claim 3, characterized in that: Step S400 includes the following steps: Step S401: Monitor the user's online behavior traces in real time, assuming that the q-th person tag B is captured in one unit period T. q The number of valid behavioral traces is n q Then, after one period T, the qth person's tag B q Total number of valid behavioral traces (n) q ) total =N q +n q In the p individual character tags, besides the character tag B q The αth character tag B, besides α The number of valid behavioral traces is n α Then, after one period T, the character tag B α Total number of valid behavioral traces (n) α ) total =N α +n α ; Step S402: When (n q ) total ≥(n α ) total If so, continue to use character tag B. q As a character's tag; when α = α1 Then tag the person As a character label.

6. A person tagging system applied to the person tagging method based on network behavior traces according to any one of claims 1-5, characterized in that: The characterization system includes a network behavior trace screening module, a risk assessment module, a network behavior warning module, and a character tag adjustment module; The network behavior trace filtering module is used to obtain the network behavior traces of users when browsing historical pages, capture keywords in the historical browsing pages, compare the keywords with the keywords in the established database to obtain normal pages and abnormal pages; and filter invalid behavior traces in the network behavior traces based on the historical browsing pages to obtain valid behavior traces. The risk assessment module is used to set several person tags containing corresponding keywords, divide the effective behavioral traces according to the set person tags, take the person tag corresponding to the most effective behavioral traces as the user's person tag, and assess the user's risk value under the person tag. The network behavior warning module is used to determine whether to designate the person tag and user as a target of attention based on the comparison result between the user's risk level and the set danger threshold, and to restrict the user's network behavior. The user tag adjustment module is used to monitor a user's online behavior traces in real time over a period of time, and adjust the user's user tags according to the total number of valid behavior traces corresponding to different user tags.

7. The character tagging system according to claim 6, characterized in that: The network behavior trace filtering module includes a network behavior trace segmentation unit and a target behavior trace filtering unit; The network behavior trace segmentation unit is used to acquire network behavior traces of users browsing historical pages, capture keywords in historical browsing pages, compare the keywords with keywords in the established database, and obtain normal pages and abnormal pages; the target behavior trace filtering unit is used to filter invalid behavior traces in network behavior traces based on historical browsing pages, and exclude invalid behavior traces to obtain valid behavior traces.

8. The character tagging system according to claim 6, characterized in that: The risk assessment module includes a person tag acquisition unit and a risk calculation unit; The user tag acquisition unit is used to set several user tags containing corresponding keywords, divide the valid behavioral traces according to the set user tags, and take the user tag corresponding to the most frequent valid behavioral traces as the user's user tag; the risk calculation unit is used to assess the user's risk value under the user tags.

Images