CN115168453A - Interest analysis method and system for e-commerce data push - Google Patents

Interest analysis method and system for e-commerce data push Download PDF

Info

Publication number
CN115168453A
CN115168453A CN202210881434.1A CN202210881434A CN115168453A CN 115168453 A CN115168453 A CN 115168453A CN 202210881434 A CN202210881434 A CN 202210881434A CN 115168453 A CN115168453 A CN 115168453A
Authority
CN
China
Prior art keywords
push
data
interest
pushing
mining data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
CN202210881434.1A
Other languages
Chinese (zh)
Inventor
赵雨
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Individual
Original Assignee
Individual
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Individual filed Critical Individual
Priority to CN202210881434.1A priority Critical patent/CN115168453A/en
Publication of CN115168453A publication Critical patent/CN115168453A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24—Querying
    • G06F16/245—Query processing
    • G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
    • G06F16/2465—Query processing support for facilitating data mining operations in structured databases
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0631—Item recommendations

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Software Systems (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • Computational Linguistics (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Health & Medical Sciences (AREA)
  • General Business, Economics & Management (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Strategic Management (AREA)
  • Marketing (AREA)
  • Economics (AREA)
  • Development Economics (AREA)
  • Fuzzy Systems (AREA)
  • Probability & Statistics with Applications (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The embodiment of the application discloses an interest analysis method and system for e-commerce data push; determining first push interest mining data, obtaining second push interest mining data corresponding to an electronic commerce interactive session to be subjected to data push analysis, then judging whether the first push interest mining data meet a set big data push requirement or not by combining the first push interest mining data and the second push interest mining data, and determining a big data push indication by combining the first push interest mining data meeting the set big data push requirement. Therefore, each determined first pushing interest mining data and each cached second pushing interest mining data can be contrastively analyzed according to the set big data pushing requirement, so that repeated pushing is reduced, the big data pushing efficiency for electronic commerce is improved, the disturbance of frequent repeated pushing on the user of the electronic commerce is avoided, and the intelligent degree of big data pushing is improved.

Description

Interest analysis method and system for e-commerce data push
Technical Field
The present application relates to the field of data processing technologies, and in particular, to an interest analysis method and system for e-commerce data push.
Background
Electronic commerce refers to all activities of online (on-line) transaction, online payment (or payment by delivery), intelligent distribution and related comprehensive services realized through contactless interaction in a network environment, and is a business operation mode which is completely innovative or simulates traditional business processes to a certain extent and is characterized by being applied in an informatization means. Electronic commerce has the characteristics of universality, convenience, integrity, safety, harmony and the like, and is widely applied under the current big data era. Data push is one of important business branches of electronic commerce, the push efficiency of the data push is concerned consistently, however, related data push technologies often have inefficient push problems such as repeated push, user annoyance, and the like.
Disclosure of Invention
An object of the present application is to provide an interest analysis method and system related to e-commerce data push.
The technical scheme of the application is realized by at least some of the following embodiments.
An interest analysis method related to electronic commerce data push, which is applied to a big data processing system, and at least comprises the following steps: acquiring a current e-commerce user behavior record collected by a first user behavior processing thread of an e-commerce interactive session to be subjected to data push analysis; determining first push interest mining data in combination with the current e-commerce user behavior record; obtaining second pushed interest mining data corresponding to the e-commerce interactive session to be subjected to data pushing analysis, wherein the second pushed interest mining data is pushed interest mining data recorded by user behaviors of a previous e-commerce of the e-commerce interactive session to be subjected to data pushing analysis, and a user behavior processing thread for collecting the user behaviors recorded by the user behaviors of the previous e-commerce comprises the first user behavior processing thread; and determining a big data pushing instruction by combining the first pushing interest mining data on the basis of determining that the first pushing interest mining data meets the set big data pushing requirement by combining the first pushing interest mining data and the second pushing interest mining data.
The method and the device for processing the electronic commerce interaction data are applied to the embodiment, and can determine first push interest mining data by combining current electronic commerce user behavior records collected by a first user behavior processing thread, obtain second push interest mining data corresponding to the electronic commerce interaction session to be subjected to data push analysis, judge whether the first push interest mining data meet a set big data push requirement or not by combining the first push interest mining data and the second push interest mining data, and determine a big data push indication by combining the first push interest mining data meeting the set big data push requirement. Therefore, each determined first pushing interest mining data and each cached second pushing interest mining data can be contrastively analyzed according to the set big data pushing requirement, so that repeated pushing is reduced, the big data pushing efficiency for electronic commerce is improved, the disturbance of frequent repeated pushing on the e-commerce users is avoided, and the intelligent degree of big data pushing is improved.
In some independent embodiments, the obtaining of the current e-commerce user behavior record collected by the first user behavior processing thread of the e-commerce interaction session to be subjected to data push analysis includes: obtaining an original e-commerce activity log collected by the first user behavior processing thread; and extracting the original electric commercial business activity log according to a set extraction step length to obtain the current electric commercial user behavior record.
The method and the device for pushing the interest mining data are applied to the embodiment, some current electric commercial user behavior records which are low in value and repeated can be filtered, the calculation cost when the first pushing interest mining data is determined by combining the current electric commercial user behavior records is reduced, and the timeliness of pushing interest analysis is improved.
In some embodiments, the method further comprises: determining a first interest event mining report meeting the set push interest mining requirement in the current electric appliance user behavior record by combining the configured AI neural network; the determining first push interest mining data in combination with the current utility user behavior record comprises: determining first push interest mining data in conjunction with the first interest event mining report.
The method and the device for mining the first interest event are applied to the embodiment, and the first interest event mining report reaching the set push interest mining requirement can be accurately and reliably determined.
In some embodiments, the first push interest-mining data includes a first distinguishing label of the first user behavior processing thread; after determining first push interest mining data in conjunction with the current utility user behavior record, the method further comprises: determining a thread cluster distinguishing label corresponding to the first user behavior processing thread by combining the first distinguishing label; caching the first pushed interest mining data to a cloud server corresponding to the thread cluster distinguishing label; the obtaining of the second pushed interest mining data corresponding to the e-commerce interactive session to be subjected to data pushing analysis includes: and obtaining second pushed interest mining data, except the first pushed interest mining data, cached in the cloud server corresponding to the thread cluster distinguishing tag.
The method and the device for processing the user behavior of the same thread cluster are applied to the embodiment, the first push interest mining data can be summarized in time according to the thread cluster, timeliness of the second push interest mining data determined by combining the user behavior processing threads of the same thread cluster is improved, and execution efficiency of the whole scheme is improved.
In some embodiments, the method further comprises determining whether the first push interest mining data meets a set big data push requirement by: determining a first electrical business requirement item covered by the first push interest mining data and a second electrical business requirement item covered by the second push interest mining data; determining that the first push interest mining data does not meet the set big data push requirement on the basis that the first electricity business requirement item and the second electricity business requirement item are the same electricity business requirement item, and the time interval between the data collection time points corresponding to the first push interest mining data and the second push interest mining data respectively does not exceed the set time; or determining that the first push interest mining data meets the set big data push requirement on the basis that the first electric business requirement item and the second electric business requirement item are not the same electric business requirement item or the basis that the first electric business requirement item and the second electric business requirement item are the same electric business requirement item and the time interval between the data collection time corresponding to the first push interest mining data and the second push interest mining data respectively exceeds a set time interval.
The embodiment can ensure that the big data pushing indication is not determined any more for the first pushing interest mining data which contains the same electric business requirement items as the second pushing interest mining data in the set time length, so that the repeated pushing frequency of the same electric business requirement items can be reduced, the pushing resources are saved, and the pushing intelligence degree is improved.
In some embodiments, the method further comprises determining whether the first electrical business requirement item and the second electrical business requirement item are the same electrical business requirement item based on: obtaining a requirement item description field of the first electrical business requirement item and a requirement item description field of the second electrical business requirement item; determining a field word vector similarity value between the demand item description field of the first electrical business demand item and the demand item description field of the second electrical business demand item, and if the field word vector similarity value exceeds a set decision value, determining that the first electrical business demand item and the second electrical business demand item are the same electrical business demand item.
The embodiment can accurately determine whether the first electric business requirement item and the second electric business requirement item are the same, and further accurately and reliably judge whether the first pushing interest mining data meets the set big data pushing requirement.
In some embodiments, the requirement item description field includes at least one of: the demand comment description vector, the service feedback emotion vector and the business operation habit vector.
When the method is applied to the above embodiment, whether the first electric business requirement item and the second electric business requirement item are the same electric business requirement item or not is judged according to the multiple requirement item description fields, so that the judgment precision and the reliability of the big data push requirement can be ensured, and the defect of repeated determination or missed determination of the big data push indication can be overcome.
In some embodiments, the method further comprises: receiving a processing request which is sent by a big data pushing system and contains a duration change interval; and adjusting the set time length by combining the change time length interval.
The big data pushing system can flexibly adjust the frequency of big data pushing and improve the intelligent degree of the whole scheme in big data pushing analysis and big data pushing decision.
In some independent embodiments, the first push interest mining data includes local behavior record content in the current e-commerce user behavior record that meets the set push interest mining requirement; the determining a big data push indication in combination with the first push interest mining data comprises: according to the local behavior record content, adding annotation knowledge in the current electric commercial user behavior record, wherein the annotation knowledge is used for highlighting the record content meeting the set pushing interest mining requirement in the current electric commercial user behavior record.
The method and the system are applied to the embodiment, the target electric business requirement items in the current electric business user behavior record can be highlighted, and subsequent targeted data mining analysis is facilitated.
In some embodiments, the first pushed interest mining data comprises pushed interest mining data of not less than one e-commerce interest topic; the method further comprises the step of determining whether the first push interest mining data meets a set big data push requirement based on the following steps: determining a first electrical business requirement item encompassed by the first push interest mining data and a second electrical business requirement item encompassed by the second push interest mining data; determining that the first push interest mining data does not meet the set big data push requirement on the basis of determining that the first electric business requirement item and the second electric business requirement item are the same electric business requirement item and determining that a first electric business interest topic of push interest mining data covered in the second push interest mining data is completely consistent with a second electric business interest topic of push interest mining data covered in the first push interest mining data; determining that the first pushed interest mining data meets the set big data pushing requirement on the basis that the first electric business requirement item and the second electric business requirement item are not the same electric business requirement item or that the first electric business requirement item and the second electric business requirement item are the same electric business requirement item and the second electric business interest topic comprises pushed interest mining data of a third electric business interest topic, wherein the third electric business interest topic is a residual electric business interest topic except the first electric business interest topic.
The method and the device for pushing the first pushed interest mining data are applied to the embodiment, the big data pushing indication is not determined any more for the first pushed interest mining data which are the same as the E-business interest topics of the pushed interest mining data contained in the second pushed interest mining data, so that the repeated pushing frequency of items with the same E-business requirement can be reduced, the pushing resources are saved, and the pushing intelligence degree is improved.
In some embodiments, the determining a big data push indication in conjunction with the first push interest mining data comprises: determining a big data pushing instruction by combining with pushing interest mining data of the third e-commerce interest topic in the first pushing interest mining data, wherein the description vectors of the big data pushing instructions determined by the pushing interest mining data of different e-commerce interest topics are different; wherein the description vector of the big data push indication at least comprises an output rule of the big data push indication and/or an output object of the big data push indication.
The method and the device are applied to the embodiment, different big data pushing instructions can be determined according to the pushing interest mining data of different E-commerce interest topics, and the big data pushing instructions are matched with personalized output rules, so that the decision auxiliary quality of the big data pushing instructions in the later pushing instruction is guaranteed.
A big data processing system, comprising: a memory for storing an executable computer program, a processor for implementing the above method when executing the executable computer program stored in the memory.
A computer-readable storage medium, on which a computer program is stored which, when executed, performs the above-described method.
Drawings
FIG. 1 is a schematic diagram illustrating one communication configuration of a large data processing system in which embodiments of the present application may be implemented.
FIG. 2 is a flow diagram illustrating an interest analysis method with respect to e-commerce data push that can implement an embodiment of the present application.
Fig. 3 is an architecture diagram illustrating an application environment in which an interest analysis method related to e-commerce data push can be implemented according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be described in further detail with reference to the attached drawings, the described embodiments should not be considered as limiting the present application, and all other embodiments obtained by a person of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application. In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the present application only and is not intended to be limiting of the application.
FIG. 1 is a block diagram illustrating one communication configuration of a big data processing system 100 that may implement embodiments of the present application, the big data processing system 100 including a memory 101 for storing executable computer programs, and a processor 102 for implementing an interest analysis method with respect to e-commerce data push in embodiments of the present application when executing the executable computer programs stored in the memory 101.
Fig. 2 is a flowchart illustrating an interest analysis method for e-commerce data push, which may implement an embodiment of the present application, and the interest analysis method for e-commerce data push may be implemented by the big data processing system 100 shown in fig. 1, and further may include the technical solutions described in the following related steps.
STEP101, obtaining the current e-commerce user behavior record collected by the first user behavior processing thread of the e-commerce interaction session to be subjected to data push analysis.
STEP102, determining a first push interest mining data in combination with the current e-commerce user behavior record.
And STEP103, acquiring second pushing interest mining data corresponding to the e-commerce interactive session to be subjected to data pushing analysis.
The second pushed interest mining data is pushed interest mining data recorded by user behaviors of a previous e-commerce of the e-commerce interactive session to be subjected to data pushing analysis, and the user behavior processing thread for collecting the user behavior records of the previous e-commerce comprises the first user behavior processing thread.
STEP104, determining a big data pushing instruction by combining the first pushing interest mining data on the basis of determining that the first pushing interest mining data reaches a set big data pushing requirement by combining the first pushing interest mining data and the second pushing interest mining data.
According to the scheme, the second push interest mining data recorded by the user behavior of the prior electric power company and the first push interest mining data recorded by the user behavior of the current electric power company are combined for joint analysis, so that whether the first push interest mining data meets the set big data push requirement or not is accurately judged, and the big data push indication is determined on the basis that the first push interest mining data meets the set big data push requirement. By combining historical data to determine the big data pushing indication, the big data pushing indication can be prevented from being frequently and repeatedly determined, and therefore the intelligent degree of pushing the auxiliary guidance information of the big data pushing indication is improved.
The following description and explanation are for example of STEPs 101 to STEPs 104, and should not be construed as limiting STEPs 101 to STEPs 104, nor should be construed as an essential feature for implementing STEPs 101 to STEPs 104.
For example, for STEP101, the e-commerce interaction session to be subjected to data push analysis may be a multi-session interaction record of a related e-commerce service, such as an existing cross-border e-commerce shopping interaction record, a metastic VR service interaction record, and the like, which is not limited herein. The electronic commerce interaction session to be subjected to data push analysis corresponds to a plurality of user behavior processing threads (used for collecting and collecting electronic commerce user behavior data to form an electronic commerce user behavior record), the first user behavior processing thread can be one of the user behavior processing threads, the first user behavior processing thread is configured according to actual conditions, the user behavior processing thread can be a behavior data acquisition program or a behavior data acquisition module, and the user behavior processing thread is authorized by a corresponding electronic commerce client before acquiring the user behavior, namely, the implementation of the whole scheme of the application is realized under the condition that the electronic commerce client/the electronic commerce user knows and agrees.
In some independent embodiments, when obtaining the current e-commerce user behavior record collected by the first user behavior processing thread of the e-commerce interaction session to be subjected to data push analysis, the current e-commerce user behavior record may be obtained by obtaining an original e-commerce activity log collected by the first user behavior processing thread, and then extracting the original e-commerce activity log according to a set extraction step size (sampling step size or sampling period).
Illustratively, the first user-activity processing thread collects multiple sets of electrical business activity logs at short steps in collecting raw electrical business activity logs, and the first user-activity processing thread may collect at decimation steps of 15 sets of electrical business activity logs collected per s. The extracting the original e-commerce activity logs according to the set extraction step length may be extracting a plurality of groups of e-commerce activity logs at set intervals, for example, extracting once at 5 intervals, and then taking the extracted e-commerce activity logs as the current e-commerce user behavior record.
Therefore, the repeated current electric commercial user behavior records with low value can be filtered, the calculation overhead when the first pushing interest mining data is determined by combining the current electric commercial user behavior records is reduced, and the timeliness of pushing interest analysis is improved.
For STEP102, in some independent embodiments, before determining the first push interest mining data in conjunction with the current electric utility user behavior record, a first interest event mining report meeting the set push interest mining requirement in the current electric utility user behavior record may be determined in conjunction with a configured AI neural network (such as an existing convolutional neural network, a deep learning network, a feature pyramid network, or the like), and then when determining the first push interest mining data in conjunction with the current electric utility user behavior record, the first push interest mining data may be determined in conjunction with the first interest event mining report.
In some examples, the set push interest mining requirement may be understood as a push interest mining requirement matching the e-commerce interaction session to be subjected to the data push analysis, for example, if the e-commerce interaction session to be subjected to the data push analysis is a metasma VR service interaction record, the set push interest mining requirement may be understood as "personal information security protection against metasma interaction", and the like.
For some possible examples, in view of the fact that the current e-commerce user behavior record covers a large amount of record content, there may be other noise content (such as some content on a flow-based and non-user behavior level) besides the e-commerce interaction session to be subjected to data push analysis, the AI neural network may first determine session data in the current e-commerce user behavior record according to a preset session information capture window, then perform target e-commerce requirement item mining on the session data in the current e-commerce user behavior record to obtain different e-commerce requirement items (such as GUI upgrading requirements, information security protection requirements, and the like), and the AI neural network may further output feature data of the target e-commerce requirement items, such as heat, type, user feedback text, and the like. Further, the mined target electric business requirement items can be selected according to the set push interest mining requirement, for example, a first interest event mining report associated with information such as "personal information safety protection against meta-universe interaction" is selected, the interest event mining report may include a plurality of electric business requirement items, and correspondingly, the push interest mining data may also include a plurality of electric business requirement items and related activity preference/tendency analysis information. Therefore, the first interest event mining report reaching the set push interest mining requirement can be accurately and reliably determined.
In some independent embodiments, after obtaining the current e-commerce user behavior record, the current e-commerce user behavior record may be sent to a cloud server, and the cloud server may determine a cache space of the current e-commerce user behavior record after receiving the current e-commerce user behavior record, and send the cache space to a user behavior analysis unit, so that a user behavior analysis unit (e.g., an e-commerce user behavior record analysis module deployed in the system) may obtain the current e-commerce user behavior record in combination with the cache space when analyzing the current e-commerce user behavior record.
In some independent embodiments, when determining the first push interest mining data in combination with the current electric utility user behavior record, the current electric utility user behavior record may be input into a configured decision tree analysis model, the decision tree analysis model performs decision analysis on the content in the current electric utility user behavior record to obtain a decision analysis result, and then determines whether the decision analysis result meets a set push interest mining requirement, for example, whether the push interest mining requirement includes set requirement items (for example, requirement items with a large influence range and a high retention contribution degree for a user) in the decision analysis result. When a decision analysis result in the decision analysis results meets the push interest mining requirement (if a set requirement exists), the first push interest mining data can be determined by combining the decision analysis results.
Further, the first push interest mining data may include a collection time of the current electric business user behavior record, a cache space of the current electric business user behavior record, a third distinguishing tag of the target electric business requirement item, feature data (requirement detail feature, requirement category feature, requirement feedback feature, and the like) of the target electric business requirement item, a first distinguishing tag (identification for distinguishing different user behavior processing threads) for obtaining the first user behavior processing thread, and the like.
Further, for STEP103, it may be performed by a mining data optimization unit. In some independent embodiments, after determining the first push interest mining data, the user behavior analysis unit may add the first push interest mining data to a candidate data set, and obtain the first push interest mining data from the candidate data set every first set time interval by a mining data optimization unit, or the user behavior analysis unit may directly send the first push interest mining data to the mining data optimization unit, and then perform STEP103 by the mining data optimization unit.
In some embodiments, the first push interest-mining data includes a first distinguishing label of the first user behavior processing thread; after the first pushed interest mining data is determined by combining the current electric appliance user behavior record, determining a thread cluster distinguishing label corresponding to the first user behavior processing thread by combining the first distinguishing label; caching the first pushed interest mining data to a cloud server corresponding to the thread cluster distinguishing label; then, when second push interest mining data corresponding to the e-commerce interaction session to be subjected to data push analysis is obtained, second push interest mining data, except the first push interest mining data, cached in the cloud server corresponding to the thread cluster distinguishing tag can be obtained.
Illustratively, a plurality of user behavior processing threads are disassembled according to thread clusters (such as thread grouping), each thread cluster may include at least one user behavior processing thread set for the same e-commerce interaction session to be subjected to data push analysis, and the user behavior processing threads belonging to the same thread cluster are used for collecting e-commerce user behavior records of the same e-commerce interaction session to be subjected to data push analysis; the first user behavior processing thread is any user behavior processing thread in any thread cluster, and each user behavior processing thread corresponds to a thread distinguishing label, such as a first distinguishing label of the first user behavior processing thread; each thread cluster corresponds to one thread cluster distinguishing label; multiple cache spaces may be included in the cloud server, and each cache space may record historical push interest mining data corresponding to different thread cluster distinguishing tags, respectively.
It can be understood that, after the first push interest mining data is received, a first distinguishing tag in the first push interest mining data may be determined, a mapping list (for example, a corresponding relationship) between a set user behavior processing thread distinguishing tag and a thread cluster distinguishing tag is combined, a thread cluster distinguishing tag matched with the first distinguishing tag is determined, the first push interest mining data is cached in a cloud server corresponding to the thread cluster distinguishing tag, and then, in order to obtain historical push interest mining data determined by a current e-commerce user behavior record collected by a user behavior processing thread (for example, of an e-commerce interaction session to be subjected to data push analysis) in the same thread cluster, when second push interest mining data corresponding to the e-commerce interaction session to be subjected to data push analysis is obtained, information except for the first push interest mining data cached in the cloud server may be used as second push interest mining data.
Based on the above content, the first push interest mining data can be summarized in time according to the thread cluster, so that the timeliness of the second push interest mining data determined by the user behavior processing thread which determines to finish the contract thread cluster later is improved, and the execution efficiency of the whole scheme is improved.
For some other design ideas, the first pushed interest mining data and the historical pushed interest mining data may be sequentially and directly recorded in a cache space according to a received time sequence, when second pushed interest mining data corresponding to an e-commerce interaction session to be subjected to data push analysis is obtained, a thread cluster distinguishing tag corresponding to the first user behavior processing thread may be determined in combination with the first distinguishing tag, and meanwhile, a thread cluster distinguishing tag corresponding to a user behavior processing thread of the historical pushed interest mining data is determined in combination with a second distinguishing tag of the historical pushed interest mining data in the cache space (for example, the thread distinguishing tag of the user behavior processing thread used for obtaining the current e-commerce user behavior record of the historical pushed interest mining data), and then the historical pushed interest mining data identical to the thread cluster distinguishing tag corresponding to the first user behavior processing thread is used as the second pushed interest mining data.
In the case of STEP104, it is also necessary to determine whether the first push interest mining data and the second push interest mining data meet the set big data push requirement before STEP104 is implemented.
In some independent embodiments, the second push interest-mining data is determined in conjunction with a second interest-event mining report (such as a record of prior e-commerce user behavior of the e-commerce interaction session to be data-push analyzed); when determining whether the first pushed interest mining data meets the set big data pushing requirement, a first electric business requirement item covered by the first pushed interest mining data and a second electric business requirement item covered by the second pushed interest mining data may be determined first.
Further, on the basis that the first electric business requirement item and the second electric business requirement item are the same electric business requirement item, and the time length interval between the data collection time moments corresponding to the first push interest mining data and the second push interest mining data respectively does not exceed a set time length, it is determined that the first push interest mining data does not meet the set big data push requirement; or, on the basis that the first electrical business requirement item and the second electrical business requirement item are not the same electrical business requirement item, or on the basis that the first electrical business requirement item and the second electrical business requirement item are the same electrical business requirement item, and a duration interval between data collection times corresponding to the first push interest mining data and the second push interest mining data respectively exceeds a set duration, determining that the first push interest mining data reaches the set big data push requirement.
It will be appreciated that if the first and second electrical business requirement profiles are the same electrical business requirement profile, indicating that an excessively large data push indication has been determined for that electrical business requirement profile, then the large data push indication is no longer determined; when the first electrical business requirement event and the second electrical business requirement event are different electrical business requirement events, it indicates that no excessively large data push indication is determined for the electrical business requirement events. And any electric business requirement item is likely to appear multiple times in a certain time period, if the time duration between the data collection time corresponding to the first push interest mining data and the second push interest mining data respectively exceeds a set time duration (such as 12 h), it is indicated that the electric business requirement item is likely to reappear (it can also be understood that the requirement corresponding to the electric business requirement item is not solved or the execution of related big data push is abnormal), so a big data push instruction should be determined again for the electric business requirement item, and when the time duration does not exceed the set time duration, the big data push instruction does not need to be determined repeatedly.
Therefore, the first pushing interest mining data which contains the same electricity business requirement items as the second pushing interest mining data in the set time length can not determine the big data pushing indication any more, so that the repeated pushing frequency of the same electricity business requirement items can be reduced, the pushing resources are saved, and the intelligent degree of pushing is improved.
Exemplary design considerations for determining whether the first push interest mining data meets the set big data push requirement may include the following described in relation to steps.
(1) And determining whether the first electrical business requirement item and the second electrical business requirement item are the same electrical business requirement item.
First, in some embodiments that may be independent, when determining whether the first electrical business requirement item and the second electrical business requirement item are the same electrical business requirement item, the requirement item description field of the first electrical business requirement item and the requirement item description field of the second electrical business requirement item may be obtained first, then the field word vector similarity value (such as cosine similarity) between the requirement item description field of the first electrical business requirement item and the requirement item description field of the second electrical business requirement item is determined, and if the field word vector similarity value exceeds the set determination value, the first electrical business requirement item and the second electrical business requirement item are determined to be the same electrical business requirement item.
For example, the requirement item description field (such as item feature information) of the first e-commerce requirement item may be mined from the current e-commerce user behavior record; the requirement item description field (such as item feature information) of the second electricity business requirement item may be cached in advance, or may be mined from the second interest event mining report.
Therefore, whether the first electric business requirement item is the same as the second electric business requirement item can be accurately determined, and whether the first pushing interest mining data meets the set big data pushing requirement or not can be accurately and reliably judged.
In some embodiments, in the mining of the demand event description field, the demand event description field of the first electrical business demand event and the demand event description field of the second electrical business demand event may be mined in combination with the configured deep residual error network. Wherein the requirement item description field comprises at least one of: demand comment description vectors (comment text information for demand items), service feedback emotion vectors (viewpoint emotion of users in e-commerce service interaction data), business operation habit vectors (user operation behavior characteristics).
For example, the deep residual network may be used to mine the requirement comment description vector and the business operation habit vector, and the feature mining rule may be used to analyze the e-commerce user behavior records collected by multiple user behavior processing threads and mine the service feedback emotion vector.
In this way, whether the first electric business requirement item and the second electric business requirement item are the same electric business requirement item or not is judged according to the multiple requirement item description fields, the judgment precision and the reliability of the big data push requirement can be ensured, and the situation of repeated determination or missed determination of the big data push indication is improved.
On the basis of the above, determining a field word vector similarity value between the requirement item description field of the first electric business requirement item and the requirement item description field of the second electric business requirement item, then comparing and analyzing the determined field word vector similarity value with a set judgment value, and when the field word vector similarity value is greater than the set judgment value, determining that the first electric business requirement item and the second electric business requirement item are the same electric business requirement item; when the field word vector similarity value is smaller than the set judgment value, determining that the first electric business requirement item and the second electric business requirement item are different electric business requirement items.
For example, if the field word vector similarity value is 0.85 and the set decision value is 0.8, the first and second electrical business requirement items are determined to be the same electrical business requirement item, and if the field word vector similarity value is 0.5 and the set decision value is 0.6, the first and second electrical business requirement items are determined to be different electrical business requirement items.
(2) And determining whether a time interval between data collection moments respectively corresponding to the first push interest mining data and the second push interest mining data exceeds a set time.
Illustratively, the first pushed interest mining data includes a first collection time, the first collection time is a time when the first user behavior processing thread collects the current e-commerce user behavior record, the second pushed interest mining data includes a second collection time, the second collection time is a time when the user behavior processing thread collects the second interest event mining report, and a difference between the first collection time and the second collection time is the duration interval. And then comparing and analyzing the duration interval with a set duration, if the duration interval exceeds the set duration, determining that the first pushed interest mining data meets a set big data pushing requirement, if the duration interval does not exceed the set duration, determining that the first pushed interest mining data does not meet the set big data pushing requirement, and then cleaning the first pushed interest mining data.
For example, if the first collection time is 15s, and the second collection time is 6s, the duration interval is 9s, if the set duration is 10s, the duration interval does not exceed the set duration, the first push interest mining data does not reach the set big data push requirement, and if the set duration is 5s, the duration interval exceeds the set duration, the first push interest mining data reaches the set big data push requirement.
For another example, for some other design considerations, it may be determined whether the time duration interval is greater than the set time duration, and then it is determined whether the first electrical business requirement item and the second electrical business requirement item are the same electrical business requirement item. For example, a duration interval between the first collection time and the second collection time may be determined, and then it is determined whether the duration interval is greater than the set duration, if so, it is determined that the first mining data of interest to be pushed reaches the set big data pushing requirement, if not, it is determined whether the first mining data of interest to be pushed and the second mining data of interest to be pushed are the same electric business requirement, if so, the first mining data of interest to be pushed does not reach the set big data pushing requirement, and if not, the first mining data of interest to be pushed reaches the set big data pushing requirement.
For other design considerations, the first pushed interest mining data includes pushed interest mining data of not less than one e-commerce interest topic, and when determining whether the first pushed interest mining data meets a set big data pushing requirement, first determining a first e-commerce requirement item covered by the first pushed interest mining data and a second e-commerce requirement item covered by the second pushed interest mining data.
Further, on the basis that the first electric business requirement item and the second electric business requirement item are determined to be the same electric business requirement item, and that a first electric business interest topic (such as category information of push interests) of push interest mining data covered in the second push interest mining data is completely consistent with a second electric business interest topic of push interest mining data covered in the first push interest mining data, determining that the first push interest mining data does not reach the set big data push requirement;
determining that the first pushed interest mining data meets the set big data pushing requirement on the basis that the first electric business requirement item and the second electric business requirement item are not the same electric business requirement item or that the first electric business requirement item and the second electric business requirement item are the same electric business requirement item and the second electric business interest topic comprises pushed interest mining data of a third electric business interest topic, wherein the third electric business interest topic is a residual electric business interest topic except the first electric business interest topic.
For example, the first push interest mining data may include push interest mining data for different e-commerce interest topics for different situations. In terms of cross-border e-commerce processing environments, the second e-commerce interest topic may be a category, detail information, attention, and the like of an e-commerce business requirement item.
It can be seen that if the first and second electric business requirement items are the same electric business requirement item and the first and second electric business interest topics are completely consistent, it indicates that push interest mining data for all electric business interest topics of the electric business requirement item has been subjected to push indication determination, so that the first push interest mining data may not be subjected to repeated push indication determination, if the first and second electric business requirement items are different, it indicates that the first and second push interest mining data are push interest mining data determined for different pedestrians, it should be subjected to push indication determination, if the first and second electric business requirement items are the same and a third electric business interest mining data included in the second electric business interest topic is included in the push interest mining data, it indicates that although the push interest mining data has been determined for the electric business requirement item, there is a difference in push interest mining data processing between the first and second electric business interest mining data, and it is determined that there is a push indication difference in push indication processing for the same push interest mining data, and thus, there is no difference in push indication processing for the push interest mining data that there is a push indication determination.
It will be appreciated that the above-described related steps may be implemented in connection with determining whether the first electrical business requirement profile and the second electrical business requirement profile are the same electrical business requirement profile. A second e-commerce interest topic of the first pushed interest mining data may be output by the AI neural network.
Therefore, the big data pushing indication is not determined any more for the first pushing interest mining data which are the same as the E-business interest topics of the pushing interest mining data contained in the second pushing interest mining data, so that the repeated pushing frequency of the same E-business requirement items can be reduced, the pushing resources are saved, and the pushing intelligence degree is improved.
In some independent embodiments, when determining a big data push indication in combination with the first push interest mining data, a big data push indication may be determined in combination with push interest mining data of the third e-commerce interest topic in the first push interest mining data, where there is a difference in description vectors of the big data push indications determined by the push interest mining data of different e-commerce interest topics; wherein the description vector of the big data push indication at least comprises an output rule of the big data push indication and/or an output object of the big data push indication.
For example, the output rule of the big data push instruction may include voice output, image-text output, and the like, and the output object of the big data push instruction may include a push system, a third-party e-commerce platform system, a third-party meta-space server, and the like.
For example, if the third e-commerce interest topic is an e-commerce interest topic, a big data push instruction may be output to the third-party e-commerce platform system to instruct the third-party e-commerce platform system to perform big data push based on the third e-commerce interest topic.
Therefore, different big data pushing instructions can be determined according to the pushing interest mining data of different E-commerce interest topics, and the big data pushing instructions are matched with personalized output rules, so that the decision auxiliary quality of the big data pushing instructions in the later pushing instruction is guaranteed.
In some independent embodiments, the big data push system may further adjust the set time duration. Exemplarily, a processing request including a duration change interval sent by a big data push system may be received; and adjusting the set time length by combining the change time length interval. For example, after receiving the processing request with the changed duration interval of 2s, the set duration may be adjusted to 2s.
Based on the method, the big data pushing system can flexibly adjust the frequency degree of big data pushing, and the intelligent degree of the whole scheme in the big data pushing analysis and big data pushing decision making is improved.
In some independent embodiments, the first push interest mining data further includes a content of a local behavior record in the current electric utility user behavior record that meets the set push interest mining requirement; when a big data push instruction is determined by combining the first push interest mining data, annotation knowledge can be added in the current electric appliance user behavior record according to the local behavior record content, and the annotation knowledge is used for highlighting the record content meeting the set push interest mining requirement in the current electric appliance user behavior record.
Therefore, the target electricity business requirement items in the current electricity business user behavior record can be highlighted, and subsequent targeted data mining analysis is facilitated.
In some independent embodiments, when determining a big data pushing instruction in combination with the first pushing interest mining data, a big data pushing instruction may be determined in combination with the local behavior record content, the feature data of the target e-commerce requirement item, the third distinguishing label of the target e-commerce requirement item, the current e-commerce user behavior record, and the like, the big data pushing instruction may include a text pushing instruction, a voice pushing instruction, a graphics context pushing instruction, and the like, and the big data pushing instruction may be sent to other sides, such as a big data pushing system, through a data transmission unit.
For example, when the mining data optimization unit sends the big data pushing instruction to the data transmission unit, the mining data optimization unit may also send the big data pushing instruction to the candidate data set first, and then the data transmission unit obtains the big data pushing instruction from the candidate data set every second set time.
The interest analysis method related to e-commerce data push provided by the embodiment of the application can determine first push interest mining data by combining current e-commerce user behavior records collected by a first user behavior processing thread, obtain second push interest mining data corresponding to e-commerce interaction sessions to be subjected to data push analysis, judge whether the first push interest mining data meets a set big data push requirement or not by combining the first push interest mining data and the second push interest mining data, and determine a big data push indication by combining the first push interest mining data meeting the set big data push requirement. Therefore, each determined first pushing interest mining data and the cached second pushing interest mining data can be contrastively analyzed according to the set big data pushing requirement, so that repeated pushing is reduced, the big data pushing efficiency for electronic commerce is improved, and the annoyance of frequent repeated pushing to the e-commerce users is avoided.
Under some independently implementable design considerations, after determining a big data push indication in conjunction with the first push interest mining data, the method may further comprise: and sending the big data push instruction to a cross-border e-commerce platform system, so that the cross-border e-commerce platform system carries out big data push processing based on the big data push instruction.
In the embodiment of the application, if the first push interest mining data relates to user interest items or requirement items of cross-border e-commerce, the determined big data push instruction can be issued to the corresponding cross-border e-commerce platform system, and then targeted big data push processing is performed through the cross-border e-commerce platform system, so that the efficiency of big data push is improved, resource waste caused by frequent repeated push is avoided, and the big data push processing is performed on the side of the cross-border e-commerce platform system, so that the operation overhead on the side of the big data processing system can be reduced, that is, the big data processing system only needs to determine the big data push instruction (assist the related platform systems to perform the big data push processing), and thus, a plurality of platform systems, such as the cross-border e-commerce platform system, a digital office platform system or a metaspace service platform system, and the like, can be docked in parallel.
Under some design ideas which can be independently implemented, enabling the cross-border e-commerce platform system to perform big data push processing based on the big data push instruction may include the following contents: enabling the cross-border e-commerce platform system to perform feedback element extraction on a first push feedback text set obtained based on the big data push instruction to obtain a first feedback element feature map set corresponding to the first push feedback text set; performing push decision analysis according to the first feedback element feature map set to obtain a second feedback element feature map set of push decision analysis; performing feature decoding processing on the second feedback element feature map set to obtain a first push decision report set corresponding to the second feedback element feature map set; and carrying out big data push processing according to the first push decision report.
In some independently implementable design considerations, after obtaining the first set of push decision reports corresponding to the second set of feedback element feature maps, the method further includes: performing push decision analysis according to at least part of push feedback texts in the first push feedback text set to obtain a second push decision report set of push decision analysis; and obtaining a third pushing decision report set of pushing decision analysis according to the first pushing decision report set and the second pushing decision report set.
Under some design ideas which can be independently implemented, the feedback element extraction is performed on a first push feedback text set obtained based on the big data push instruction, so as to obtain a first feedback element feature map set corresponding to the first push feedback text set, including: performing word vector extraction on a first push feedback text set obtained based on the big data push indication to obtain a first word vector distribution set corresponding to the first push feedback text set; and performing feature mapping on the first word vector distribution set to obtain a first feedback element feature map set corresponding to the first push feedback text set.
Under some design ideas that can be independently implemented, performing push decision analysis according to the first feedback element feature diagram set to obtain a second feedback element feature diagram set of push decision analysis, including: and inputting the first feedback element feature graph set into a first natural language processing model, and pushing decision analysis through the first natural language processing model to obtain a second feedback element feature graph set corresponding to the first feedback element feature graph set.
Under some design ideas that can be independently implemented, the performing feature decoding processing on the second feedback element feature map set to obtain a first push decision report set corresponding to the second feedback element feature map set includes: performing feature decoding processing on the second feedback element feature map set to obtain a second word vector distribution set corresponding to the second feedback element feature map set; and performing inverse word vector extraction on the second word vector distribution set to obtain a first push decision report set corresponding to the second feedback element feature map set.
Under some design ideas which can be independently implemented, performing feature decoding processing on the second feedback element feature map set to obtain a second word vector distribution set corresponding to the second feedback element feature map set, including: and circularly processing the second feedback element feature map set by adopting a trigger unit relu to obtain a second word vector distribution set corresponding to the second feedback element feature map set.
Under some design considerations that can be implemented independently, performing push decision analysis according to at least a part of the push feedback texts in the first push feedback text set to obtain a second push decision report set of push decision analysis includes: inputting at least part of push feedback texts in the first push feedback text set into a second natural language processing model, and obtaining a second push decision report set corresponding to the at least part of push feedback texts through push decision analysis of the second natural language processing model.
Under some design ideas which can be independently implemented, the at least part of the push feedback texts include X push feedback texts obtained in real time from the first push feedback text set, where X is a positive integer, and the number of the push feedback texts in the first push feedback text set is greater than or equal to X.
In some design considerations that can be implemented independently, the obtaining a third set of push decision reports for push decision analysis according to the first set of push decision reports and the second set of push decision reports includes: performing decision vector mining on the first push decision report set to obtain a first decision vector corresponding to the first push decision report set; performing decision vector mining on the second push decision report set to obtain a second decision vector corresponding to the second push decision report set; performing first splicing processing according to the first decision vector and the second decision vector to obtain a first splicing decision vector; and obtaining a third pushing decision report set of pushing decision analysis according to the first splicing decision vector.
Under some design ideas that can be implemented independently, obtaining, according to the first concatenation decision vector, a third set of push decision reports for push decision analysis includes: performing regression analysis processing on the first splicing decision vector to obtain a regression analysis decision vector; and obtaining a third push decision report set of push decision analysis according to the first splicing decision vector and the regression analysis decision vector.
Under some independently implementable design considerations, the first stitching decision vector comprises a plurality of layers; the obtaining a third push decision report set of push decision analysis according to the first splicing decision vector and the regression analysis decision vector comprises: performing second splicing processing on the first splicing decision vector of the last layer to obtain a second splicing decision vector; and obtaining a third push decision report set of push decision analysis according to the second splicing decision vector and the regression analysis decision vector.
Under some independently implementable design considerations, the first stitching decision vector comprises a plurality of layers; performing regression analysis processing on the first splicing decision vector to obtain a regression analysis decision vector, including: and performing regression analysis processing on the first splicing decision vector of the first layer to obtain a regression analysis decision vector.
In some design ideas that can be implemented independently, a decision vector thereof is that the performing decision vector mining on the first push decision report set to obtain a first decision vector corresponding to the first push decision report set includes: performing multi-layer decision vector mining on the first push decision report set to obtain a multi-layer first decision vector corresponding to the first push decision report set; the performing decision vector mining on the second push decision report set to obtain a second decision vector corresponding to the second push decision report set includes: performing multi-layer decision vector mining on the second push decision report set to obtain a multi-layer second decision vector corresponding to the second push decision report set; performing a first splicing process according to the first decision vector and the second decision vector to obtain a first splicing decision vector, including: and for any layer in the plurality of layers, performing decision vector splicing according to the first decision vector of the layer and the second decision vector of the layer to obtain a first splicing decision vector of the layer.
Under some design ideas that can be implemented independently, performing decision vector stitching according to the first decision vector of the layer and the second decision vector of the layer to obtain a first stitching decision vector of the layer includes: in response to that the layer does not belong to the last layer, performing decision vector splicing on the first decision vector of the layer, the second decision vector of the layer and the first splicing decision vector of the next layer of the layer to obtain a first splicing decision vector of the layer; and/or, in response to the layer belonging to the last layer, performing decision vector splicing on the first decision vector of the layer and the second decision vector of the layer to obtain a first splicing decision vector of the layer.
Based on the above technical scheme, feedback element extraction is performed on a first push feedback text set obtained based on the big data push instruction to obtain a first feedback element feature map set corresponding to the first push feedback text set, push decision analysis is performed according to the first feedback element feature map set to obtain a second feedback element feature map set of push decision analysis, feature decoding is performed on the second feedback element feature map set to obtain a first push decision report set corresponding to the second feedback element feature map set, so that the first feedback element feature map set obtained based on feedback element extraction (feedback feature mining) is subjected to push decision analysis, the resource cost in the push decision analysis process is not large, the push decision analysis timeliness is high, and the push decision analysis efficiency can be improved to perform big data push quickly and accurately.
Fig. 3 is an architecture diagram illustrating an application environment of an interest analysis method for e-commerce data push in which a big data processing system 100 and an e-commerce client 200 communicating with each other may be included, in which an embodiment of the present application may be implemented. Based on this, the big data processing system 100 and the e-commerce client 200 implement or partially implement an interest analysis method related to e-commerce data push of the embodiment of the present application when running.
The embodiments of the present application have been described above with reference to the accompanying drawings, and have at least the following beneficial effects: the method comprises the steps of determining first push interest mining data by combining current electric business user behavior records collected by a first user behavior processing thread, obtaining second push interest mining data corresponding to an electronic business interaction session to be subjected to data push analysis, judging whether the first push interest mining data meet a set big data push requirement or not by combining the first push interest mining data and the second push interest mining data, and determining a big data push indication by combining the first push interest mining data meeting the set big data push requirement. Therefore, each determined first pushing interest mining data and each cached second pushing interest mining data can be contrastively analyzed according to the set big data pushing requirement, so that repeated pushing is reduced, the big data pushing efficiency for electronic commerce is improved, the disturbance of frequent repeated pushing on the e-commerce users is avoided, and the intelligent degree of big data pushing is improved.
The above description is only a preferred embodiment of the present application, and is not intended to limit the scope of the present application.

Claims (10)

1. An interest analysis method related to e-commerce data push, which is applied to a big data processing system, and at least comprises the following steps:
acquiring a current e-commerce user behavior record collected by a first user behavior processing thread of an e-commerce interactive session to be subjected to data push analysis; determining first push interest mining data by combining the current e-commerce user behavior record, and obtaining second push interest mining data corresponding to the e-commerce interactive session to be subjected to data push analysis; wherein: the second pushed interest mining data is pushed interest mining data recorded by user behaviors of a previous electric company of the electronic commerce interactive session to be subjected to data pushing analysis, and the user behavior processing threads for collecting the user behavior records of the previous electric company comprise the first user behavior processing threads;
and determining a big data pushing instruction by combining the first pushing interest mining data on the basis of determining that the first pushing interest mining data reaches a set big data pushing requirement by combining the first pushing interest mining data and the second pushing interest mining data.
2. The method of claim 1, wherein obtaining a record of current e-commerce user behavior collected by a first user behavior processing thread of an e-commerce interaction session to be subjected to data push analysis comprises:
obtaining an original e-commerce activity log collected by the first user behavior processing thread;
and extracting the original electric commercial business activity log according to a set extraction step length to obtain the current electric commercial user behavior record.
3. The method of claim 1, further comprising: determining a first interest event mining report meeting the set push interest mining requirement in the current electric commercial user behavior record by combining the configured AI neural network;
the determining first push interest mining data in combination with the current utility user behavior record comprises: determining first push interest mining data in conjunction with the first interest event mining report.
4. The method of claim 1, wherein the first push interest mining data comprises a first distinguishing label for the first user behavior processing thread;
after determining a first push interest mining data in conjunction with the current utility user behavioral record, the method further comprises: determining a thread cluster distinguishing label corresponding to the first user behavior processing thread by combining the first distinguishing label; caching the first pushed interest mining data to a cloud server corresponding to the thread cluster distinguishing label;
the obtaining of the second pushed interest mining data corresponding to the e-commerce interactive session to be subjected to data pushing analysis includes: and obtaining second pushed interest mining data, except the first pushed interest mining data, cached in the cloud server corresponding to the thread cluster distinguishing tag.
5. The method of claim 1, further comprising determining whether the first push interest mining data meets a set big data push requirement by:
determining a first electrical business requirement item encompassed by the first push interest mining data and a second electrical business requirement item encompassed by the second push interest mining data;
determining that the first push interest mining data does not meet the set big data push requirement on the basis that the first electricity business requirement item and the second electricity business requirement item are the same electricity business requirement item, and the time interval between the data collection time points corresponding to the first push interest mining data and the second push interest mining data respectively does not exceed the set time;
determining that the first push interest mining data meets the set big data push requirement on the basis that the first electricity business requirement item and the second electricity business requirement item are not the same electricity business requirement item or that the first electricity business requirement item and the second electricity business requirement item are the same electricity business requirement item and that a duration interval between data collection moments corresponding to the first push interest mining data and the second push interest mining data respectively exceeds a set duration.
6. The method of claim 5, further comprising determining whether the first electrical business requirement item and the second electrical business requirement item are the same electrical business requirement item based on:
obtaining a requirement item description field of the first electrical business requirement item and a requirement item description field of the second electrical business requirement item;
determining a field word vector similarity value between the demand item description field of the first electrical business demand item and the demand item description field of the second electrical business demand item, and if the field word vector similarity value exceeds a set decision value, determining that the first electrical business demand item and the second electrical business demand item are the same electrical business demand item.
7. The method of claim 6, wherein the requirement item description field comprises at least one of: a demand comment description vector, a service feedback emotion vector and a business operation habit vector;
wherein the method further comprises: receiving a processing request which is sent by a big data pushing system and contains a duration change interval; and adjusting the set time length by combining the change time length interval.
8. The method of claim 7, wherein the first push interest mining data comprises local behavioral record content in the current utility user behavioral record that meets the set push interest mining requirement;
the determining a big data push indication in combination with the first push interest mining data comprises: according to the local behavior record content, adding annotation knowledge in the current electric commercial user behavior record, wherein the annotation knowledge is used for highlighting the record content meeting the set pushing interest mining requirement in the current electric commercial user behavior record.
9. The method of claim 1, wherein the first pushed interest mining data comprises pushed interest mining data of not less than one e-commerce interest topic; the method further comprises the step of determining whether the first push interest mining data meets a set big data push requirement based on the following steps:
determining a first electrical business requirement item covered by the first push interest mining data and a second electrical business requirement item covered by the second push interest mining data;
determining that the first push interest mining data does not meet the set big data push requirement on the basis of determining that the first electric business requirement item and the second electric business requirement item are the same electric business requirement item and determining that a first electric business interest topic of push interest mining data covered in the second push interest mining data is completely consistent with a second electric business interest topic of push interest mining data covered in the first push interest mining data;
determining that the first pushed interest mining data meets the set big data pushing requirement on the basis that the first electric business requirement item and the second electric business requirement item are not the same electric business requirement item or that the first electric business requirement item and the second electric business requirement item are the same electric business requirement item and the second electric business interest topic contains pushed interest mining data of a third electric business interest topic, wherein the third electric business interest topic is a residual electric business interest topic except the first electric business interest topic;
wherein the determining a big data push indication in combination with the first push interest mining data comprises: determining a big data pushing instruction by combining with pushing interest mining data of the third e-commerce interest topic in the first pushing interest mining data, wherein the description vectors of the big data pushing instructions determined by the pushing interest mining data of different e-commerce interest topics are different; wherein the description vector of the big data push indication at least comprises an output rule of the big data push indication and/or an output object of the big data push indication.
10. A big data processing system, comprising:
a memory for storing an executable computer program, a processor for implementing the method of any one of claims 1-9 when executing the executable computer program stored in the memory.
CN202210881434.1A 2022-07-26 2022-07-26 Interest analysis method and system for e-commerce data push Withdrawn CN115168453A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210881434.1A CN115168453A (en) 2022-07-26 2022-07-26 Interest analysis method and system for e-commerce data push

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210881434.1A CN115168453A (en) 2022-07-26 2022-07-26 Interest analysis method and system for e-commerce data push

Publications (1)

Publication Number Publication Date
CN115168453A true CN115168453A (en) 2022-10-11

Family

ID=83496698

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210881434.1A Withdrawn CN115168453A (en) 2022-07-26 2022-07-26 Interest analysis method and system for e-commerce data push

Country Status (1)

Country Link
CN (1) CN115168453A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115860836A (en) * 2022-12-07 2023-03-28 广东南粤分享汇控股有限公司 E-commerce service pushing method and system based on user behavior big data analysis
CN115905702A (en) * 2022-12-06 2023-04-04 鄄城县馨宁网络科技有限公司 Data recommendation method and system based on user demand analysis
CN117668368A (en) * 2023-12-18 2024-03-08 重庆机电职业技术大学 E-commerce data pushing method and system based on big data

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115905702A (en) * 2022-12-06 2023-04-04 鄄城县馨宁网络科技有限公司 Data recommendation method and system based on user demand analysis
CN115905702B (en) * 2022-12-06 2023-10-10 雨果跨境(厦门)科技有限公司 Data recommendation method and system based on user demand analysis
CN115860836A (en) * 2022-12-07 2023-03-28 广东南粤分享汇控股有限公司 E-commerce service pushing method and system based on user behavior big data analysis
CN115860836B (en) * 2022-12-07 2023-09-26 广东南粤分享汇控股有限公司 E-commerce service pushing method and system based on user behavior big data analysis
CN117668368A (en) * 2023-12-18 2024-03-08 重庆机电职业技术大学 E-commerce data pushing method and system based on big data

Similar Documents

Publication Publication Date Title
CN115168453A (en) Interest analysis method and system for e-commerce data push
US9590880B2 (en) Dynamic collection analysis and reporting of telemetry data
CN110399925A (en) Risk Identification Method, device and the storage medium of account
CN108363811A (en) Device identification method and device, electronic equipment, storage medium
CN114840486B (en) User behavior data acquisition method and system and cloud platform
CN110442712B (en) Risk determination method, risk determination device, server and text examination system
US20100268776A1 (en) System and Method for Determining Information Reliability
US20160012074A1 (en) System and method for providing contextual analytics data
CN103440199A (en) Method and device for guiding test
CN113486983A (en) Big data office information analysis method and system for anti-fraud processing
CN113158069A (en) Interactive topic scene analysis method based on big data, server and medium
CN111917848A (en) Data processing method based on edge computing and cloud computing cooperation and cloud server
CN116225848A (en) Log monitoring method, device, equipment and medium
CN113377640B (en) Method, medium, device and computing equipment for explaining model under business scene
CN110569424A (en) Information recommendation method and device
CN111611483B (en) Object portrait construction method, device and equipment and storage medium
CN112905892A (en) Big data processing method and big data server applied to user portrait mining
CN111143688A (en) Evaluation method and system based on mobile news client
Costa et al. A three level sensor ranking method based on active perception
Raman et al. Mutual Information Pre-processing Based Broken-stick Linear Regression Technique for Web User Behaviour Pattern Mining.
CN108449425A (en) A kind of monitoring system of machine and its method based on Internet of Things
CN116707834B (en) Distributed big data evidence obtaining and analyzing platform based on cloud storage
US20230121911A1 (en) Methods and systems for determining severity of disruptions in communication networks in non-homogenous environments
CN118195446A (en) Logistics package interception method, device and storage medium
CN114297234A (en) Method and device for identifying key behavior data

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WW01 Invention patent application withdrawn after publication

Application publication date: 20221011

WW01 Invention patent application withdrawn after publication