Personnel correlation analysis method, device, electronic equipment and storage medium

Abstract

The embodiment of the present invention provides a personnel association analysis method, which obtains each portrait event and each code event to be processed; based on each portrait event and each code event, obtains a person-code integration index; based on the person-code integration index, obtains a person-code fit; based on the person-code fit, obtains a correlation analysis result between each portrait event and each code event. The present invention also discloses a personnel association analysis device, an electronic device, and a computer-readable storage medium. By combining and analyzing portrait events and code events, a more accurate correlation analysis result can be determined without a large number of portrait events and code events, thereby reducing the consumption of analysis resources, shortening the analysis time, and improving the analysis efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of event analysis, and in particular to a personnel correlation analysis method, device, electronic equipment and storage medium. Background Art

[0002] At present, in the field of event analysis, the person associated with the target event is determined by performing correlation analysis on the portrait event or code event corresponding to the target event. However, in the existing method, the technician manually performs correlation analysis on the portrait event to obtain the correlation analysis results of multiple target persons, or performs correlation analysis on the code event to obtain the correlation analysis results of multiple target persons. When analyzing the portrait event and the code event separately, more events are required to obtain more accurate correlation analysis results. Analyzing more events will consume more computing resources and computing time, making the efficiency of correlation analysis low. Summary of the invention

[0003] The embodiment of the present invention provides a method for analyzing personnel association, and obtains the person-code fitting degree through the person-code fitting degree; based on the person-code fitting degree, obtains the association analysis results of multiple target persons, and does not require a large number of portrait events and code events to determine a more accurate association analysis result, thereby reducing the analysis resource consumption, shortening the analysis time, and improving the analysis efficiency.

[0004] In a first aspect, an embodiment of the present invention provides a method for analyzing personnel association, comprising the following steps:

[0005] Get each portrait event and each code event to be processed;

[0006] Based on each of the portrait events and each of the code events, obtaining a person-code integration index;

[0007] Based on the human-code integration index, obtaining the human-code fitting degree;

[0008] Based on the person-code fitting degree, a correlation analysis result between each of the person portrait events and each of the code events is obtained.

[0009] Optionally, before obtaining the human-code fitting degree based on the human-code integration index, the method further includes:

[0010] Based on each of the portrait events and the code events, obtaining a person-code association degree, a person-code time dispersion degree, and a person-code behavior similarity degree;

[0011] The obtaining of the human-code fitting degree based on the human-code integration index includes:

[0012] The human-code fitting degree is obtained based on the human-code integration index, the human-code association degree, the human-code time dispersion degree and the human-code behavior similarity.

[0013] Optionally, the step of obtaining a person-code integration index based on each of the portrait events and each of the code events includes:

[0014] Based on the time information of each portrait event and the time information of each code event, obtaining the number of times the portrait and code are combined;

[0015] Determining a preset smoothing coefficient based on each of the portrait events, each of the code events and a preset analysis purpose;

[0016] Calculate the number of pre-processing human-code merges using the number of human-code merges and the preset smoothing coefficient;

[0017] The human-code merge index is calculated based on the pre-processed human-code merge times, the preset smoothing coefficient and the smoothing coefficient.

[0018] Optionally, based on each of the portrait events and each of the code events, obtaining a person-code association degree includes:

[0019] Based on the number of times the portrait and code are merged, the number of single merges of each portrait event and each code event, the number of codes merged with each portrait event, and the number of portraits merged with each code event are obtained;

[0020] Based on the single bounce number and the code quantity, obtaining a bounce frequency;

[0021] Based on the total number of portraits in each portrait event and the number of portraits, obtaining a merge frequency index;

[0022] The human-code association degree is obtained based on the merge frequency and the merge frequency index.

[0023] Optionally, based on each of the portrait events and each of the code events, obtaining a person-code time discreteness includes:

[0024] Based on the total number of portraits in each portrait event and the number of times the portrait code is merged, the number of days for the portrait code to be merged is obtained;

[0025] Based on the number of days for human code integration, the total number of days for human code integration is obtained;

[0026] The human code time dispersion is calculated using the preset merge days threshold, the human code merge days and the human code merge total days.

[0027] Optionally, the step of obtaining the similarity of person-code behavior based on each of the portrait events and each of the code events includes:

[0028] Using each of the portrait events, construct a plurality of first person activity vectors corresponding to the plurality of target persons, where one target person corresponds to one first person activity vector;

[0029] Using each of the code events, construct multiple second person activity vectors corresponding to the multiple target persons, one target person corresponds to one second person activity vector;

[0030] The cosine similarity is calculated for each of the first person activity vectors and each of the second person activity vectors to obtain the person-code behavior similarity.

[0031] Optionally, before the step of obtaining the human-code fitting degree based on the human-code alignment index, the human-code association degree, the human-code time dispersion degree and the human-code behavior similarity, the method further includes:

[0032] Obtaining a first weight of the human code association, a second weight of the human code time dispersion, and a third weight of the human code behavior similarity;

[0033] The step of obtaining the human-code fitting degree based on the human-code integration index, the human-code association degree, the human-code time dispersion degree and the human-code behavior similarity comprises:

[0034] Obtaining an intermediate score based on the person-code association degree, the person-code time dispersion degree, the person-code behavior similarity degree, the first weight, the second weight, and the third weight;

[0035] The human-code fitting degree is calculated using the human-code fitting index and the intermediate score.

[0036] In a second aspect, an embodiment of the present invention provides a device for analyzing personnel association, the device comprising:

[0037] An acquisition module, used for acquiring various portrait events and various code events to be processed;

[0038] A first obtaining module, used for obtaining a person-code merging index based on each of the portrait events and each of the code events;

[0039] A second obtaining module is used to obtain a human-code fitting degree based on the human-code integration index;

[0040] The third obtaining module is used to obtain the correlation analysis result between each of the portrait events and each of the code events based on the person-code fitting degree.

[0041] In a third aspect, an embodiment of the present invention provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps in the personnel association analysis method provided in the embodiment of the present invention when executing the computer program.

[0042] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps in the method for analyzing the degree of association of personnel provided in the embodiment of the invention are implemented.

[0043] In the embodiment of the present invention, each portrait event and each code event to be processed are obtained; based on each portrait event and each code event, a person-code integration index is obtained; based on the person-code integration index, a person-code fit is obtained; based on the person-code fit, a correlation analysis result between each portrait event and each code event is obtained. By combining and analyzing portrait events and code events to obtain a person-code integration index, and based on the person-code integration index, a final correlation analysis result is obtained, and a large number of portrait events and code events are not required to determine a more accurate correlation analysis result, thereby reducing the consumption of analysis resources, shortening the analysis time, and improving the analysis efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0045] Figure 1 is a flow chart of a method for analyzing personnel relevance provided by an embodiment of the present invention;

[0046] Figure 2 is a flow chart of another personnel association analysis method provided by an embodiment of the present invention;

[0047] Figure 3 It is a structural schematic diagram of a personnel association analysis device provided by an embodiment of the present invention;

[0048] Figure 4 is a schematic structural diagram of another personnel association analysis device provided by an embodiment of the present invention;

[0049] Figure 5 is a structural schematic diagram of a first obtaining module provided by an embodiment of the present invention;

[0050] Figure 6is a structural schematic diagram of a fourth obtaining module provided by an embodiment of the present invention;

[0051] Figure 7 is a structural schematic diagram of another fourth obtaining module provided in an embodiment of the present invention;

[0052] Figure 8 is a structural schematic diagram of another fourth obtaining module provided in an embodiment of the present invention;

[0053] Fig. 9 is a schematic structural diagram of another personnel association analysis device provided by an embodiment of the present invention;

[0054] Fig.10 It is a structural schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] See also Figure 1 , Figure 1 is a flow chart of a personnel association analysis method provided by an embodiment of the present invention, such as Figure 1 As shown, the personnel correlation analysis method includes the following steps:

[0057] 101. Obtain portrait events and code events to be processed.

[0058] In an embodiment of the present invention, each portrait event and each code event to be processed are each portrait event and code event corresponding to the target user. The above-mentioned target person refers to the person who needs to be analyzed for correlation, and there needs to be multiple target persons who need to be analyzed for correlation. Each target person will correspond to a portrait event and a code event. Among them, a portrait event of a target person refers to a set of information such as the face image, body image and face video of the target person, and a code event of a target person refers to the code information of the target person. The code information may include an ID code, a QR code, an identity code, etc. Usually, a code information represents the identity of a target person, and the code information serves as an identifier of the identity of the target person.

[0059] Typically, the electronic device receives portrait events of multiple target persons and code events of multiple target persons entered by a user, and then continues to receive analysis instructions sent by the user. Based on the received analysis instructions, the electronic device performs the steps of this application using the received portrait events of multiple target persons and code events of multiple target persons.

[0060] 102. Obtain a person-code integration index based on each of the portrait events and each of the code events.

[0061] The person-code merging refers to the simultaneous appearance of a person portrait in a person portrait event and the code information in a code event at a certain time T (which can be set by the user based on demand, such as 10s or 100s, etc.) and a certain space S (which can be set by the user based on demand, such as 100m or 500m, etc.).

[0062] The person-code merging refers to the simultaneous appearance of a person portrait in a person portrait event and the code information in a code event at a certain time T (which can be set by the user based on demand, such as 10s or 100s, etc.) and a certain space S (which can be set by the user based on demand, such as 100m or 500m, etc.).

[0063] The number of times of person-code merging refers to the number of times a certain person in the person event and the code information in the code event appear simultaneously at a certain time T and a certain space S.

[0064] The person-code integration index is obtained based on the number of person-code integration times. The more person-code integration times are, the higher the person-code integration index of the portrait and the code information is. When the number of integration times reaches a certain level, the impact on the person-code fit will tend to be flat, and the person-code integration index will also tend to be flat.

[0065] The person-code association degree is used to evaluate the importance of a certain code information to a certain person portrait.

[0066] The time dispersion of person code refers to the fact that compared with the intermittent person code merging, the time dispersion of person code and code of continuous person code merging is higher. The time dispersion of person code is an indicator reflecting the continuity of person code merging.

[0067] The person-code behavior similarity refers to the similarity of the number of times the person's portrait and the code information appear at the same time and in the same space. In the present invention, the person-code behavior similarity of the multiple target persons refers to the similarity between the person's portrait and the code information of each target person.

[0068] 103. Based on the human-code integration index, obtain the human-code fit degree

[0069] Among them, the person-code fitting degree refers to the degree of match between the person and the code based on the portrait event and the code event through a number of algorithmic analysis and integration rules, which is expressed by a specific fitting degree.

[0070] Furthermore, before obtaining the person-code fitting degree based on the person-code integration index, the method also includes: obtaining the person-code association degree, the person-code time dispersion degree and the person-code behavior similarity based on each of the portrait events and each of the code events; obtaining the person-code fitting degree based on the person-code integration index includes: obtaining the person-code fitting degree based on the person-code integration index, the person-code association degree, the person-code time dispersion degree and the person-code behavior similarity.

[0071] After obtaining the human-code alignment index, the human-code association, the human-code time dispersion and the human-code behavior similarity, the human-code fit is calculated using a preset floor(a, b) function, where floor(a, b) takes b decimal places for a, for example, floor(3.1415926, 3) = 3.141.

[0072] Specifically, before the step of obtaining the human code fitting degree based on the human code integration index, the human code association degree, the human code time dispersion degree and the human code behavior similarity, the method also includes: obtaining the first weight of the human code association degree, the second weight of the human code time dispersion degree and the third weight of the human code behavior similarity; accordingly, the step of obtaining the human code fitting degree based on the human code integration index, the human code association degree, the human code time dispersion degree and the human code behavior similarity includes: obtaining an intermediate score based on the human code association degree, the human code time dispersion degree, the human code behavior similarity, the first weight, the second weight and the third weight; and calculating the human code fitting degree using the human code integration index and the intermediate score.

[0073] It should be noted that, in the present invention, the first weight, the second weight and the third weight can be weights set by the user based on demand, and the present invention does not impose specific restrictions. Usually, based on the data saturation of portrait events and code events and the preset analysis purpose of user needs (including the refinement of the analysis results and the accuracy of the analysis results required by the user, the refinement of the analysis results and the accuracy of the analysis results are different, then the specific values ​​of the first weight, the second weight and the third weight will be different).

[0074] Specifically, the human-code fitting degree is calculated according to Formula 1, which is as follows:

[0075] C=S1W1+S2W2+S3W3

[0076] P = floor(C + A, ε)

[0077] Among them, S1 is the human code association, W1 is the first weight, S2 is the human code time dispersion, W2 is the second weight, S3 is the human code behavior similarity score, W3 is the third weight, C is the intermediate score, P is the human code fit, A is the human code integration index, ε is the number of decimal places to be retained, ε is usually 6. The floor function has the same meaning as the function described above.

[0078] Based on different situations, different weight values ​​are set to ensure that the obtained person-code fitting degree is more consistent with the actual situation, thereby improving the accuracy of the person-code fitting degree.

[0079] 104. Based on the person-code fitting degree, obtain a correlation analysis result between each of the person portrait events and each of the code events.

[0080] Based on the respective person code fits of the multiple target persons, the correlation analysis results of the multiple target persons are determined. By using the respective person code fits of the multiple target persons, it is determined whether the multiple target persons have a correlation relationship and how the correlation degree of the multiple target persons is.

[0081] Furthermore, based on the person-code fitting degree, the correlation analysis result between each of the portrait events and each of the code events is obtained, including: normalizing the person-code fitting degree to obtain a result score, and based on the result score, obtaining the correlation analysis result between each of the portrait events and each of the code events.

[0082] It should be noted that the actual value of the human-code fit may be distributed more dispersedly in a larger numerical range. At this time, all the human-code fits need to be normalized so that the corresponding result scores of all the human-code fits after normalization are distributed in the interval [0,1].

[0083] Specifically, the human-code fitting degree is normalized using Formula 2, which is:

[0084]

[0085] Among them, P i is the person-code fitting degree of the ith person in the person-code fitting degree, Q i is the result score of the i-th person, A max is the maximum value of the human code integration index, P max is the maximum value of the person code fitting degree, τ is the number of decimal places retained, τ can be 6, the meaning of the floor function refers to the above, and the i-th person is any one of the multiple target persons.

[0086] The person-code fit is normalized to make the relationship between the obtained result scores clearer, avoid the person-code fit values ​​being too dispersed, and make it difficult to determine the relevance of the target person, thus reducing the difficulty of analyzing the relevance of the target person.

[0087] In an embodiment of the present invention, portrait events of multiple target persons and code events of multiple target persons are obtained; based on the portrait events and the code events, a person-code integration index, a person-code correlation, a person-code time dispersion and a person-code behavior similarity are obtained; based on the person-code integration index, the person-code correlation, the person-code time dispersion and the person-code behavior similarity, a person-code fitting degree is obtained; based on the person-code fitting degree, a correlation analysis result of multiple target persons is obtained.

[0088] In the existing methods, a technician manually performs a correlation analysis on a portrait event to obtain the correlation analysis results of multiple target persons, or performs a correlation analysis on a code event to obtain the correlation analysis results of multiple target persons. When analyzing portrait events and code events separately, more events are required to obtain more accurate correlation analysis results. Analyzing more events will consume more computing resources and computing time, making the efficiency of correlation analysis low. In the present invention, by combining and analyzing portrait events and code events to obtain the final fit, based on the person-code fit, the correlation analysis results of multiple target persons are obtained. It does not require a large number of portrait events and code events to determine a more accurate correlation analysis result, thereby reducing the consumption of analysis resources, shortening the analysis time, and improving the analysis efficiency.

[0089] Optional, see Figure 2 , Figure 2 is a flow chart of another personnel association analysis method provided by an embodiment of the present invention, such as Figure 2 As shown, the specific steps include:

[0090] 201. Obtain the number of times portrait and code are merged based on the time information of each portrait event and the time information of each code event.

[0091] 202. Determine a preset smoothing coefficient based on each of the portrait events, each of the code events, and a preset analysis purpose.

[0092] 203. Calculate the number of pre-processing human-code merges by using the number of human-code merges and the preset smoothing coefficient.

[0093] 204. Calculate the human-code merge index based on the pre-processed human-code merge times, the preset smoothing coefficient and the smoothing coefficient.

[0094] In the embodiment of the present invention, it is necessary to count the number of times a person and a code are merged and the number of days for the person and the code to be merged based on the portrait event and the code event. The number of times a person and a code are merged may include the number of times a certain portrait is merged with a certain code information, the number of code information merged with a certain portrait, and the number of portraits merged with a certain code information, etc. The number of days for the person and a code to be merged may include the number of days for the person and a code information to be merged.

[0095] It should be noted that in the present application, a preset smoothing coefficient is determined based on the data saturation of each portrait event, the data saturation of each code event and a preset analysis purpose (the same as the preset analysis purpose described above), and the preset smoothing coefficient is a value between 0-1.

[0096] Specifically, the calculation formula for the number of pre-processing human code merges is Formula 3, as follows:

[0097]

[0098] Among them, β is the smoothing coefficient, T1 is the number of pre-processing human code merges, and T0 is the number of human code merges. Then, after obtaining the number of pre-processing human code merges, the human code merge index is obtained using formula 5, which is as follows:

[0099] Z = floor(sigmoid(T1,β),α)

[0100] Wherein, α is the number of decimal places to be retained, which is usually 6, the meaning of the floor function is as described above, and Z is the human code parallel index. In some embodiments, the pseudo code corresponding to the process of obtaining the human code parallel index is as follows:

[0101]

[0102] Human-code integration index = floor(sigmoid(number of human-code integrations after preprocessing, smoothing coefficient), retain decimal places).

[0103] The number of times of human-code merger is preprocessed. For different numbers of human-code merger, different preprocessing numbers of human-code merger are determined, so that the final human-code merger index is more in line with the analysis requirements of the correlation degree, avoiding the direct use of the number of human-code merger without processing, thereby avoiding the low accuracy of the obtained human-code merger index.

[0104] Specifically, based on each of the portrait events and each of the code events, obtaining the person-code association degree includes: based on the number of person-code merges, obtaining the number of single merges of each of the portrait events with each of the code events, the number of codes merged with each of the portrait events, and the number of portraits merged with each of the code events; based on the number of single merges and the number of codes, obtaining the merge frequency; based on the total number of portraits in each of the portrait events and the number of portraits, obtaining the merge frequency index; based on the merge frequency and the merge frequency index, obtaining the person-code association degree. Based on the portrait events, the total number of portraits is counted to obtain the person-code association degree.

[0105] Among them, the number of single merges of each portrait event and each code event (the number of single merges refers to the number of times a portrait event is merged with a code event, which can also be understood as the number of times a portrait is merged with a code information), the number of codes merged with each portrait event (the number of code events that have a merge relationship with a portrait event, which can be understood as the number of code information merged with a portrait), and the number of portraits merged with each code event (the number of portrait events that have a merge relationship with a code event, which can be understood as the number of portraits merged with a code information) can be determined from the number of person-code merges.

[0106] Among them, based on the single merge times and the code quantity, obtaining the merge frequency means: for a certain portrait E and a certain code event F, the single merge times of the portrait E and the code event F are compared with the code quantity of all code events that have a merge relationship with the portrait E, and the ratio of the ratio is the merge frequency.

[0107] Wherein, based on the total number of portraits and the number of portraits, obtaining the merge frequency index means: for a certain code event G, determining the number of portraits with a merge relationship with the code event G, then adding 1 to the number of portraits corresponding to the code event G to obtain an intermediate parameter, then calculating the ratio of the total number of portraits of all portrait events of multiple target persons to the intermediate parameter, and then using the ratio as the input of the lg function, and obtaining the output of the lg function as the merge frequency index, that is, the merge frequency index = lg intermediate parameter. Finally, the merge frequency and the merge frequency index are multiplied to obtain the final person-code association degree.

[0108] By using the above method to determine the person-code association degree, the person-code association degree can more effectively reflect the association degree between the person image and the code, thereby improving the accuracy of the person-code association degree.

[0109] Specifically, based on each of the portrait events and each of the code events, the person-code time dispersion is obtained, including: based on the total number of portraits in each of the portrait events and the number of person-code mergers, the number of person-code merger days is obtained; based on the number of person-code merger days, the total number of person-code merger days is obtained; using a preset merger day threshold, the person-code merger days and the total number of person-code merger days, the person-code merger days are calculated.

[0110] Based on the above scheme, after obtaining the number of days since the human code was merged, the human code time dispersion is continued. The steps for obtaining the human code time dispersion are as follows: based on the number of days since the human code was merged, the total number of days since the human code was merged is obtained; using the preset merged days threshold, the number of days since the human code was merged and the total number of days since the human code was merged, the human code time dispersion is calculated.

[0111] The total number of days for human code integration is the sum of the number of days for human code integration. The preset integration day threshold can be a value set by the user based on demand, such as 2 days. If the number of days for human code integration is less than 2, the human code time dispersion of the human code integration day is determined to be 0. If the number of days for human code integration is greater than or equal to 2, the ratio of the human code integration day to the total number of days for human code integration is used as the human code time dispersion.

[0112] In some embodiments, the human code time discreteness can also be solved using the following pseudo code:

[0113] Human code time dispersion = if (human code integration days < 2) { 0.0

[0115] }else{

[0116] sigmoid(set of consecutive days of human-code integration.map(days=>{Math.pow(2, Math.pow(days–1, 0.8))}).sum / Math.pow(actual total number of days of human-code integration, 0.5), 18.0)

[0117] }

[0118] It can be seen that in the above pseudo code, some other parameters are also involved. The user can adjust the corresponding parameters based on the needs, and the present invention does not specifically limit the parameter values.

[0119] For different human code integration days, different methods are used to obtain the corresponding human code time discreteness, avoiding additional analysis of invalid data with fewer human code integration days, thereby improving analysis efficiency. At the same time, the invalid data with fewer integration days are determined to be 0, avoiding inaccurate impact parameters caused by invalid data with fewer integration days that are not 0, thereby improving the accuracy of human code time discreteness.

[0120] Specifically, the step of obtaining the person-code behavior similarity based on each of the portrait events and each of the code events includes: using each of the portrait events to construct multiple first person activity vectors corresponding to multiple target persons, one target person corresponds to one first person activity vector; using each of the code events to construct multiple second person activity vectors corresponding to multiple target persons, one target person corresponds to one second person activity vector; calculating the cosine similarity of each of the first person activity vector and each of the second person activity vector to obtain the person-code behavior similarity.

[0121] Based on the above description, the person-code behavior similarity refers to the similarity of the number of times a portrait and a code appear at the same time and in the same space. The present invention solves the cosine similarity based on the first person activity vector corresponding to the portrait event and the second person activity vector corresponding to the code event to obtain the person-code behavior similarity. One target person corresponds to one person-code behavior similarity.

[0122] Specifically, the number of portrait events and code events of all target persons are counted based on the time period and the area range, so as to construct the first person activity vector and the second person activity vector respectively. Usually, the time period is divided into hours, for example, a day is divided into 24 time periods, and the area range is divided into longitude and latitude (divided by the geohash algorithm, where the precision is 7 bits).

[0123] Then, continue to calculate the human-code behavior similarity according to Formula 4, which is as follows:

[0124]

[0125] Among them, S3 is the human code behavior similarity, A is the first person's activity vector, B is the second person's activity vector, and A i is the value of the i-th dimension in A, B i is the value of the i-th dimension in B.

[0126] The cosine similarity method is used to determine the similarity between the person and the code. The similarity between the person and the code more accurately reflects the similarity information between the person and the code, which makes the final person-code fitting degree have a higher accuracy rate.

[0127] The present invention can be used in scenarios such as restoration analysis and danger warning of relevant departments. Based on the person-code fitting degree, the identity of the person associated with the event is found through the space-time accompanying behavior of the person code or the space-time position co-occurrence relationship of the person code. Finally, the individual's identity, trajectory and other characteristic data are used to unravel and locate the relevant personnel.

[0128] It should be noted that the personnel association analysis method provided in the embodiment of the present invention can be applied to devices such as smart phones, computers, servers, etc. that can perform data query.

[0129] Optional, see Figure 3 , Figure 3 is a schematic diagram of the structure of a personnel association analysis device provided by an embodiment of the present invention, such as Figure 3 As shown, the device comprises:

[0130] An acquisition module 301 is used to acquire portrait events and code events to be processed;

[0131] A first obtaining module 302 is used to obtain a person-code merging index based on each of the portrait events and each of the code events;

[0132] The second obtaining module 303 is used to obtain the human-code fitting degree based on the human-code integration index;

[0133] The third obtaining module 304 is used to obtain the correlation analysis result between each of the portrait events and each of the code events based on the person-code fitting degree.

[0134] Optional, such as Figure 4 As shown, the device also includes:

[0135] A fourth obtaining module 305 is used to obtain a person-code association degree, a person-code time dispersion degree and a person-code behavior similarity degree based on each of the person portrait events and each of the code events;

[0136] Correspondingly, the second obtaining module 303 is used to obtain the human-code fitting degree based on the human-code integration index, the human-code association degree, the human-code time dispersion degree and the human-code behavior similarity.

[0137] Optional, such as Figure 5 As shown, the first obtaining module 302 includes:

[0138] The first obtaining unit 3021 is used to obtain the number of times of combining the portrait and the code based on the time information of each portrait event and the time information of each code event;

[0139] A second obtaining unit 3022 is used to determine a preset smoothing coefficient based on each of the portrait events, each of the code events and a preset analysis purpose;

[0140] The first calculation unit 3023 is used to calculate the number of pre-processing human-code merge times by using the number of human-code merge times and the preset smoothing coefficient;

[0141] The second calculation unit 3024 is used to calculate the human-code merge index based on the pre-processed human-code merge times, the preset smoothing coefficient and the smoothing coefficient.

[0142] Optional, such as Figure 6 As shown, the fourth obtaining module 305 includes:

[0143] The third obtaining unit 3051 is used to obtain the number of single merges of each portrait event with each code event, the number of codes merged with each portrait event, and the number of portraits merged with each code event based on the number of person-code merges;

[0144] A fourth obtaining unit 3052, configured to obtain a bounce frequency based on the single bounce number and the code quantity;

[0145] A fifth obtaining unit 3053, configured to obtain a merge frequency index based on the total number of portraits in each portrait event and the number of portraits;

[0146] The sixth obtaining unit 3054 is configured to obtain the human-code association degree based on the merge frequency and the merge frequency index.

[0147] Optional, such as Figure 7 As shown, the fourth obtaining module 305 further includes:

[0148] The seventh obtaining unit 3055 is used to obtain the number of days for person-code merging based on the total number of person-images in each of the person-image events and the number of person-code merging times;

[0149] An eighth obtaining unit 3056 is used to obtain the total number of days for human code integration based on the number of days for human code integration;

[0150] The ninth obtaining unit 3057 is used to calculate the time dispersion of the human code by using the preset merge days threshold, the human code merge days and the total human code merge days.

[0151] Optional, such as Figure 8 As shown, the fourth obtaining module 305 further includes:

[0152] A first constructing unit 3058 is used to construct a plurality of first person activity vectors corresponding to a plurality of target persons by using the portrait events, where one target person corresponds to one first person activity vector;

[0153] The second construction unit 3059 is used to construct a plurality of second personnel activity vectors corresponding to a plurality of target personnel using each of the code events, wherein one target person corresponds to one second personnel activity vector;

[0154] The third calculation unit 30510 is used to calculate the cosine similarity of each of the first person activity vectors and each of the second person activity vectors to obtain the person-code behavior similarity.

[0155] Optional, such as Fig. 9 As shown, the device also includes:

[0156] A weight acquisition module 306, used to acquire a first weight of the person-code association, a second weight of the person-code time dispersion, and a third weight of the person-code behavior similarity;

[0157] Correspondingly, the second acquisition module 303 is used to obtain an intermediate score based on the human-code association degree, the human-code time dispersion, the human-code behavior similarity, the first weight, the second weight and the third weight; and calculate the human-code fitting degree using the human-code integration index and the intermediate score.

[0158] It should be noted that the personnel association analysis device provided in the embodiment of the present invention can be applied to smart phones, computers, servers and other devices that can perform business analysis at the map level.

[0159] The personnel association analysis device provided in the embodiment of the present invention can implement each process implemented by the personnel association analysis method in the above method embodiment, and can achieve the same beneficial effect. To avoid repetition, it will not be described here.

[0160] See also Fig.10 , Fig.10 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, such as Fig.10 As shown, it includes: a memory 1002, a processor 1001, and a computer program of a personnel association analysis method stored in the memory 1002 and executable on the processor 1001, wherein:

[0161] The processor 1001 is used to call the computer program stored in the memory 1002 and execute the following steps:

[0162] Get each portrait event and each code event to be processed;

[0163] Based on each of the portrait events and each of the code events, obtaining a person-code integration index;

[0164] Based on the human-code integration index, obtaining the human-code fitting degree;

[0165] Based on the person-code fitting degree, a correlation analysis result between each of the person portrait events and each of the code events is obtained.

[0166] Optionally, before the processor 1001 obtains the human-code fit degree based on the human-code integration index, the method further includes:

[0167] Based on each of the portrait events and the code events, obtaining a person-code association degree, a person-code time dispersion degree, and a person-code behavior similarity degree;

[0168] Accordingly, the processor 1001 executes the steps of obtaining the human-code fitting degree based on the human-code fitting index, including:

[0169] The human-code fitting degree is obtained based on the human-code integration index, the human-code association degree, the human-code time dispersion degree and the human-code behavior similarity.

[0170] Optionally, the processor 1001 obtains a person-code merge index based on each of the portrait events and each of the code events, including:

[0171] Based on the time information of each portrait event and the time information of each code event, obtaining the number of times the portrait and code are combined;

[0172] Determining a preset smoothing coefficient based on each of the portrait events, each of the code events and a preset analysis purpose;

[0173] Calculate the number of pre-processing human-code merges using the number of human-code merges and the preset smoothing coefficient;

[0174] The human-code merge index is calculated based on the pre-processed human-code merge times, the preset smoothing coefficient and the smoothing coefficient.

[0175] Optionally, the processor 1001 obtains a person-code association degree based on each of the portrait events and each of the code events, including:

[0176] Based on the number of times the portrait and code are merged, the number of single merges of each portrait event and each code event, the number of codes merged with each portrait event, and the number of portraits merged with each code event are obtained;

[0177] Based on the single bounce number and the code quantity, obtaining a bounce frequency;

[0178] Based on the total number of portraits in each portrait event and the number of portraits, obtaining a merge frequency index;

[0179] Based on the merge frequency and the merge frequency index, the human-code association degree is obtained.

[0180] Optionally, the processor 1001 obtains the time dispersion of the person and the code based on each of the person portrait events and each of the code events, including:

[0181] Based on the total number of portraits in each portrait event and the number of times the portrait code is merged, the number of days for the portrait code to be merged is obtained;

[0182] Based on the number of days for human code integration, the total number of days for human code integration is obtained;

[0183] The human code time dispersion is calculated using the preset merge days threshold, the human code merge days and the human code merge total days.

[0184] Optionally, the processor 1001 executes the step of obtaining the similarity of the person-code behavior based on each of the portrait events and each of the code events, including:

[0185] Using each of the portrait events, construct a plurality of first person activity vectors corresponding to the plurality of target persons, where one target person corresponds to one first person activity vector;

[0186] Using each of the code events, construct multiple second person activity vectors corresponding to the multiple target persons, one target person corresponds to one second person activity vector;

[0187] The cosine similarity is calculated for each of the first person activity vectors and each of the second person activity vectors to obtain the person-code behavior similarity.

[0188] Optionally, before the processor 1001 executes the process of obtaining the human-code fitting degree based on the human-code integration index, the human-code association degree, the human-code time dispersion degree and the human-code behavior similarity, the method further includes:

[0189] Obtaining a first weight of the human code association, a second weight of the human code time dispersion, and a third weight of the human code behavior similarity;

[0190] Accordingly, the processor 1001 executes the steps of obtaining the human-code fitting degree based on the human-code integration index, the human-code association degree, the human-code time dispersion degree and the human-code behavior similarity, including:

[0191] Obtaining an intermediate score based on the person-code association degree, the person-code time dispersion degree, the person-code behavior similarity degree, the first weight, the second weight, and the third weight;

[0192] The human-code fitting degree is calculated using the human-code fitting index and the intermediate score.

[0193] It should be noted that the electronic device provided in the embodiment of the present invention can be applied to devices such as smart phones, computers, servers, etc. that can perform personnel association analysis.

[0194] The electronic device provided in the embodiment of the present invention can implement each process implemented by the method for analyzing the personnel association degree in the above method embodiment, and can achieve the same beneficial effect. To avoid repetition, it will not be described here.

[0195] An embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the various processes of the personnel correlation analysis method or the application-side personnel correlation analysis method provided in the embodiment of the present invention are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0196] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM).

[0197] The above disclosure is only the preferred embodiment of the present invention, which certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A personnel correlation analysis method, characterized in that: The following steps are involved: Obtaining each portrait event and each code event to be processed, wherein the code event refers to the code information of the target person, and the code information serves as an identifier of the identity of the target person; Based on the time information of each portrait event and the time information of each code event, obtaining the number of times the portrait and code are combined; Determining a preset smoothing coefficient based on each of the portrait events, each of the code events and a preset analysis purpose; Calculate the number of pre-processing human-code merges using the number of human-code merges and the preset smoothing coefficient; Calculating the human-code merge index based on the pre-processed human-code merge times, the preset smoothing coefficient and the smoothing coefficient; Based on each of the portrait events and the code events, obtaining a person-code association degree, a person-code time dispersion degree, and a person-code behavior similarity degree; The person-code association degree is used to evaluate the importance of the code information to the person's portrait; the person-code time dispersion is an indicator reflecting the continuity of the person-code integration; The similarity between human-code behavior refers to the similarity of the number of times the human image and code information appear at the same time and in the same space; Based on the human-code integration index, the human-code association degree, the human-code time dispersion degree and the human-code behavior similarity degree, obtaining the human-code fitting degree; Based on the person-code fitting degree, a correlation analysis result between each of the person portrait events and each of the code events is obtained.

2. The method according to claim 1, characterized in that Based on each of the portrait events and each of the code events, obtaining a person-code association degree includes: Based on the number of times the portrait and code are merged, the number of single merges of each portrait event and each code event, the number of codes merged with each portrait event, and the number of portraits merged with each code event are obtained; Based on the single bounce number and the code quantity, obtaining a bounce frequency; Based on the total number of portraits in each portrait event and the number of portraits, obtaining a merge frequency index; The human-code association degree is obtained based on the merge frequency and the merge frequency index.

3. The method according to claim 1, characterized in that Based on each of the portrait events and each of the code events, obtaining a person-code time discreteness includes: Based on the total number of portraits in each portrait event and the number of times the portrait code is merged, the number of days for the portrait code to be merged is obtained; Based on the number of days for human code integration, the total number of days for human code integration is obtained; The human code time dispersion is calculated using the preset merge days threshold, the human code merge days and the human code merge total days.

4. The method according to claim 1, characterized in that The step of obtaining the similarity of the person-code behavior based on each of the person-portrait events and each of the code events comprises: Using each of the portrait events, construct a plurality of first person activity vectors corresponding to the plurality of target persons, where one target person corresponds to one first person activity vector; Using each of the code events, construct multiple second person activity vectors corresponding to the multiple target persons, one target person corresponds to one second person activity vector; The cosine similarity is calculated for each of the first person activity vectors and each of the second person activity vectors to obtain the person-code behavior similarity.

5. The method according to any one of claims 2 to 4, characterized in that: Before the step of obtaining the human-code fitting degree based on the human-code integration index, the human-code association degree, the human-code time dispersion degree and the human-code behavior similarity, the method further includes: Obtaining a first weight of the human code association, a second weight of the human code time dispersion, and a third weight of the human code behavior similarity; The step of obtaining the human-code fitting degree based on the human-code integration index, the human-code association degree, the human-code time dispersion degree and the human-code behavior similarity comprises: Obtaining an intermediate score based on the person-code association degree, the person-code time dispersion degree, the person-code behavior similarity degree, the first weight, the second weight, and the third weight; The human-code fitting degree is calculated using the human-code fitting index and the intermediate score.

6. A personnel association analysis device, characterized in that: The device includes: An acquisition module, used to acquire each portrait event and each code event to be processed, wherein the code event refers to the code information of the target person, and the code information serves as an identification of the identity of the target person; A first obtaining module, for obtaining the number of times the portrait and code are combined based on the time information of each portrait event and the time information of each code event; Determining a preset smoothing coefficient based on each of the portrait events, each of the code events and a preset analysis purpose; Calculate the number of pre-processing human-code merges using the number of human-code merges and the preset smoothing coefficient; Calculating the human-code merge index based on the pre-processed human-code merge times, the preset smoothing coefficient and the smoothing coefficient; A second obtaining module is used to obtain a person-code association degree, a person-code time dispersion degree and a person-code behavior similarity degree based on each of the portrait events and each of the code events; The person-code association degree is used to evaluate the importance of the code information to the person's portrait; the person-code time dispersion is an indicator reflecting the continuity of the person-code integration; The similarity between human-code behavior refers to the similarity of the number of times the human image and code information appear at the same time and in the same space; Based on the human-code integration index, the human-code association degree, the human-code time dispersion degree and the human-code behavior similarity degree, obtaining the human-code fitting degree; The third obtaining module is used to obtain the correlation analysis result between each of the portrait events and each of the code events based on the person-code fitting degree.

7. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps in the method for analyzing the degree of association of personnel as claimed in any one of claims 1 to 5 when executing the computer program.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the method for analyzing the personnel association degree according to any one of claims 1 to 5 are implemented.

Images