Search

CN-120257167-B - Marketing campaign data analysis management system and method based on two-dimension code

CN120257167BCN 120257167 BCN120257167 BCN 120257167BCN-120257167-B

Abstract

The invention discloses a marketing campaign data analysis management system and method based on two-dimension codes, and relates to the technical field of data analysis, wherein the management method comprises the following steps of establishing a behavior data management system to collect scanning data and user behavior data and generating corresponding scanning records; the method comprises the steps of expanding multidimensional evaluation on any scanning record, carrying out anomaly identification, analyzing user behavior data of any scanning record, extracting to obtain a user behavior set, randomly selecting two scanning records with different anomaly identification results, extracting anomaly characteristics from the user behavior set, analyzing influence values of different difference characteristics, constructing an anomaly evaluation model, setting an anomaly evaluation threshold value for whether anomalies exist, collecting the scanning data of a user and the user behavior data in real time, generating a real-time scanning record, and carrying out anomaly identification on the real-time scanning record based on the influence values of different difference characteristics in the real-time scanning record.

Inventors

  • LIU MI
  • CHEN JIAPING
  • Liang Kanglun

Assignees

  • 广东中源数字化技术有限公司

Dates

Publication Date
20260508
Application Date
20250407

Claims (9)

  1. 1. A marketing campaign data analysis and management method based on two-dimension codes is characterized by comprising the following steps: Step 100, establishing a behavior data management system to collect the scanning data generated by scanning the two-dimensional code each time and the user behavior data and generate corresponding scanning records; Step 200, analyzing the user behavior data of any scan record, extracting the operation behavior of the user to obtain a user behavior set, randomly selecting two scan records with different abnormal recognition results, and extracting abnormal characteristics from the user behavior set based on the difference condition between the two scan records; Step 300, analyzing the influence value of each difference feature influencing the abnormal recognition result for any abnormal scan record, constructing an abnormal evaluation model based on the deviation condition of the network data corresponding to each difference feature and the user behavior data, and setting an abnormal evaluation threshold value for any abnormal scan record; Step 400, collecting scanning data and user behavior data generated by user scanning in real time, generating a real-time scanning record and extracting difference features; the step S300 includes the steps of: Step 301, selecting an abnormal feature set of an abnormal record at will, selecting an abnormal feature from the abnormal feature set to set the abnormal feature as a target feature, counting the number of records containing the target feature in all the scan records as N and the number of abnormal records containing the target feature as N1, and calculating to obtain the abnormal occurrence frequency of the target feature as f=N1/N; Step S302, obtaining an abnormal value E of a corresponding dimension of the target feature in the selected abnormal record, setting the number of dimension features contained in the corresponding dimension as m, calculating to obtain an influence total value Y= (E-E th )/m of the target feature, wherein E th is an abnormal threshold value, setting scanning data or user behavior data of the corresponding dimension as D, obtaining scanning data or user behavior data of the corresponding dimension in each normal record, carrying out average calculation to obtain reference data D ave , calculating to obtain a deviation amplitude R= |D-D ave |/D ave of the corresponding dimension, setting the target feature as an ith dimension feature in the corresponding dimension, obtaining a deviation amplitude of the ith dimension feature as p i = R/m, and calculating to obtain an influence value of the ith dimension feature as Y i =Y×p i /f i , wherein f i is an abnormal occurrence frequency of the ith dimension feature; Step S303, randomly selecting a scanning record, extracting the influence value of each dimension feature in each dimension in the selected scanning record, wherein if one dimension feature is not a difference feature, the influence value of the dimension feature is 0, acquiring the time interval between the selected scanning record and the previous scanning record in the same scanning log as deltat, setting the time interval threshold as deltat th , if deltat is less than deltat th , setting the feature time coefficient tt=1 of the selected scanning record, otherwise, tt=0, constructing an anomaly evaluation model: ; Wherein j is a positive integer and j is E (1, a), a is the total number of features of all dimension features in the selected scanning record, y j is the influence value of the j-th dimension feature in all dimension features, and f j is the abnormal occurrence frequency of the j-th dimension feature; Step S304, obtaining abnormal evaluation values of each scanning record, obtaining a normal record section (A1 min ,A1 max ) and an abnormal record section (A2 min ,A2 max ) according to the normal record and the abnormal record respectively, setting an abnormal evaluation threshold value AEV th =A1 max if A1 max ≤A2 min , and setting an abnormal evaluation threshold value AEV th =A2 min if A1 max ＞A2 min .
  2. 2. The method for analysis and management of marketing campaign data based on two-dimensional codes according to claim 1, wherein the step S100 comprises the steps of: Step S101, when a user scans a two-dimensional code, acquiring equipment information of the user after the user agrees to acquire the equipment information, and generating a corresponding scanning log by taking the equipment information as an identification object; acquiring a time point when a user scans a two-dimensional code, collecting network data generated in the scanning process, generating scanning data, and collecting user behavior data of the user on a target page after the user enters the target page; Step S102, randomly selecting a scanning log of equipment information, sequencing each scanning record in the scanning log according to the sequence of scanning time points to obtain the time interval between any two adjacent scanning records, presetting a time interval threshold, setting the next scanning record in the two scanning records as an abnormal record if the time interval between the two scanning records is lower than the time interval threshold, and rejecting the abnormal record from all the scanning records to obtain a first record set; Step 103, randomly selecting one scanning record from the first record set, dividing network data and user behavior data stored in the selected scanning record into a plurality of dimensions according to data types, presetting a corresponding abnormal evaluation rule for each dimension to obtain an abnormal value of each dimension, and setting the selected scanning record as an abnormal record if the abnormal value of one dimension is greater than or equal to a preset abnormal threshold value.
  3. 3. The method for analysis and management of marketing campaign data based on two-dimensional codes according to claim 2, wherein the step S200 comprises the steps of: Step S201, randomly selecting a scanning record, extracting user behavior data in the scanning record, dividing the user behavior data into a plurality of data sets according to divided dimensions, and extracting each dimension to obtain a plurality of dimension characteristics; Step S202, a behavior database is built in advance, the behavior database comprises a plurality of user behaviors, a dimension is selected to correspond to a plurality of dimension characteristics at will, any one dimension characteristic is compared with any one user behavior through a preset association degree comparison rule, and the association degree between the dimension characteristic and the user behavior is calculated; selecting one dimension feature with the largest association degree and user behaviors, extracting association degrees of the selected user behaviors and other dimension features of the selected dimension, summing all the extracted association degrees, averaging to obtain average association degrees, and taking the selected user behaviors as one user behavior of the selected scanning record if the average association degrees exceed a preset association degree threshold; Step 203, matching each dimension with a corresponding user behavior, and if the average association degree of a plurality of dimension features corresponding to one dimension is lower than an association degree threshold, setting a plurality of dimension features with the average association degree lower than the association degree threshold as potential feature sets; Step S204, setting the scanning record which is not the abnormal record as the normal record, randomly selecting one normal record and one abnormal record respectively, comparing the user behavior set of the abnormal record with the user behavior set of the normal record, setting the different user behaviors as the abnormal behaviors if the user behaviors which are different from the user behavior set of the normal record exist in the user behavior set of the abnormal record, setting dimension characteristics corresponding to the abnormal behaviors as the abnormal characteristics, and obtaining the abnormal characteristic set of the selected abnormal record.
  4. 4. The method for analysis and management of marketing campaign data based on two-dimensional codes according to claim 3, wherein the step S400 comprises the steps of: step S401, each time a user scans a two-dimensional code in real time, generating a real-time scanning record in real time, and extracting difference features of scanning data and user behavior data in the real-time scanning record to obtain a plurality of difference features; Step S402, extracting a scanning log corresponding to user equipment information, obtaining a time interval between the real-time scanning record and a previous scanning record as delta t now , setting a time interval threshold as delta t th , and obtaining a characteristic time coefficient tt now =1 of the real-time scanning record if delta t now ＜Δt th , otherwise tt now =0; Step S403, obtaining the influence value of each difference characteristic, inputting the influence value into an anomaly evaluation model for accumulation to obtain an anomaly evaluation value AEV now of the real-time scanning record, setting an anomaly evaluation threshold value as AEV th , and if AEV now ＞AEV th , carrying out anomaly early warning on the real-time scanning record.
  5. 5. A marketing campaign data analysis management system for executing the two-dimensional code-based marketing campaign data analysis management method of any one of claims 1 to 4, characterized in that the management system comprises an anomaly scanning identification module, a behavior feature analysis module, an anomaly identification setting module and a real-time anomaly analysis module; the abnormal scanning identification module is used for establishing a behavior data management system to collect scanning data and user behavior data generated by scanning the two-dimensional code each time and generate corresponding scanning records; The behavior feature analysis module is used for analyzing the user behavior data of any scanning record and extracting the operation behaviors of the user to obtain a user behavior set, randomly selecting two scanning records with different abnormal recognition results, and extracting abnormal features from the user behavior set based on the difference condition between the two scanning records; The anomaly identification setting module is used for analyzing the influence value of each difference characteristic influencing the anomaly identification result for any abnormal scanning record, constructing an anomaly evaluation model based on the deviation condition of the network data corresponding to each difference characteristic and the user behavior data, and setting an anomaly evaluation threshold value for any anomaly in any scanning record; The real-time anomaly analysis module is used for collecting the scanning data generated by the scanning of the user and the user behavior data in real time, generating a real-time scanning record and extracting the difference characteristics, and identifying the anomaly of the real-time scanning record based on the influence value of each difference characteristic in the real-time scanning record.
  6. 6. The marketing campaign data analysis management system of claim 5, wherein the anomaly scanning recognition module comprises a behavioral data collection unit and an anomaly record recognition unit; The behavior data acquisition unit is used for establishing a behavior data management system to acquire the scanning data generated by scanning the two-dimensional code each time and the user behavior data and generate corresponding scanning records, and the abnormal record identification unit is used for analyzing the influence values of different characteristics affecting the abnormal identification result for any abnormal scanning record.
  7. 7. The marketing campaign data analysis management system of claim 5, wherein the behavioral characteristics analysis module comprises a user behavioral analysis unit and an anomaly characteristic extraction unit; The abnormal feature extraction unit is used for arbitrarily selecting two scan records with different abnormal recognition results, and extracting abnormal features from the user behavior set based on the difference condition between the two scan records.
  8. 8. The marketing campaign data analysis management system of claim 5, wherein the anomaly identification setting module comprises a feature impact analysis unit and an assessment model construction unit; The characteristic influence analysis unit is used for analyzing influence values of various difference characteristics influencing an abnormal recognition result for any abnormal scanning record, and the evaluation model construction unit is used for constructing an abnormal evaluation model to set an abnormal evaluation threshold value for any abnormal scanning record based on deviation conditions of network data corresponding to various difference characteristics and user behavior data.
  9. 9. The marketing campaign data analysis management system of claim 5, wherein the real-time anomaly analysis module comprises a real-time scanning acquisition unit and an anomaly change identification unit; The system comprises a real-time scanning acquisition unit, an abnormal change identification unit and a real-time scanning record generation unit, wherein the real-time scanning acquisition unit is used for establishing a behavior data management system to acquire scanning data generated by scanning a two-dimensional code each time and user behavior data and generate corresponding scanning records, and the abnormal change identification unit is used for carrying out abnormal identification on the real-time scanning records based on the influence values of different features in the real-time scanning records.

Description

Marketing campaign data analysis management system and method based on two-dimension code Technical Field The invention relates to the technical field of data analysis, in particular to a marketing campaign data analysis management system and method based on two-dimension codes. Background The traditional two-dimension code marketing system links users with marketing contents mainly through static two-dimension codes and combines public numbers of enterprises to realize user interaction, and the users can acquire various information only by scanning the codes through smart phones, so that interaction between the users and brands can be improved; Although the prior art has been advanced and has advanced to a certain extent, the two-dimensional code marketing still has significant defects in data analysis and processing, and because the marketer can only analyze the data received by the background, invalid code scanning behaviors such as test code scanning, malicious form swiping and real user data are mixed, key index distortion is caused, and in data processing, only basic indexes such as scanning times are relied on, and deep analysis cannot be performed by combining with multidimensional behavior paths, so that policy accuracy is affected. Disclosure of Invention The invention aims to provide a marketing campaign data analysis management system and method based on two-dimension codes, which are used for solving the problems in the prior art. In order to achieve the above purpose, the invention provides a marketing campaign data analysis management method based on two-dimension codes, which comprises the following steps: Step 100, establishing a behavior data management system to collect the scanning data generated by scanning the two-dimensional code each time and the user behavior data and generate corresponding scanning records; Step 200, analyzing the user behavior data of any scan record, extracting the operation behavior of the user to obtain a user behavior set, randomly selecting two scan records with different abnormal recognition results, and extracting abnormal characteristics from the user behavior set based on the difference condition between the two scan records; Step 300, analyzing the influence value of each difference feature influencing the abnormal recognition result for any abnormal scan record, constructing an abnormal evaluation model based on the deviation condition of the network data corresponding to each difference feature and the user behavior data, and setting an abnormal evaluation threshold value for any abnormal scan record; Step 400, collecting scanning data and user behavior data generated by user scanning in real time, generating a real-time scanning record and extracting difference features, and identifying the abnormality of the real-time scanning record based on the influence value of each difference feature in the real-time scanning record. Further, step S100 includes the steps of: Step S101, acquiring equipment information of a user after the user agrees to acquire the equipment information when the user scans the two-dimensional code, generating a corresponding scanning log by taking the equipment information as an identification object, acquiring a time point when the user scans the two-dimensional code, acquiring network data generated in the scanning process, generating scanning data, acquiring user behavior data of the user on a target page after the user enters the target page, summarizing the scanning data of the user and the user behavior data to obtain a scanning record corresponding to one-time scanning of the user, wherein the network data generated in the scanning process comprises data of the two-dimensional code, user equipment information, position information, jump route links, permission calling and the like; Step S102, randomly selecting a scanning log of equipment information, sequencing each scanning record in the scanning log according to the sequence of scanning code time points to obtain a time interval between any two adjacent scanning records, presetting a time interval threshold, setting the next scanning record in the two scanning records as an abnormal record if the time interval between the two scanning records is lower than the time interval threshold, and eliminating the abnormal record from all the scanning records to obtain a first record set, wherein the time interval can directly reflect whether the scanning code is invalid or not, and the reliability of the generated related data is low because the scanning code behavior exceeds the normal scanning code frequency; step 103, randomly selecting a scanning record from the first record set, dividing network data and user behavior data stored in the selected scanning record into a plurality of dimensions according to data types, presetting a corresponding abnormality evaluation rule for each dimension to obtain an abnormality value of each dimension, and if the abnormality value of one d