Metacosm user behavior analysis closed-loop method and system based on big data and medium
By collecting, preprocessing and analyzing the multimodal data of Metaverse users, generating user behavior profiles and performing funnel model decomposition, the problem of data integration and value conversion faults in Metaverse user behavior analysis is solved, and closed-loop optimization of the entire process is achieved.
Patent Information
- Application Number
- CN202511301192.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-12
- Publication Date
- 2025-10-17
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing user behavior analysis methods are unable to adapt to the metaverse scenario, resulting in a gap between data integration and value transformation, a single analysis dimension, and an inability to form a complete closed loop of data collection, analysis, and application.
By collecting multimodal data, performing preprocessing and data bit correction, generating user behavior profile data, and utilizing funnel model decomposition and user tracing and grouping sorting, we generate business process funnel reports and user grouping and value analysis tables, and combine behavioral prediction models to optimize business plans.
It realizes closed-loop analysis of metaverse user behavior, adapts to scenario characteristics, realizes multi-source data integration and full-process value transformation, and optimizes virtual scenarios, operation strategies and product functions.
Smart Images

Figure CN120804610A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of metaverse user behavior analysis, in particular to a metaverse user behavior analysis closed loop method, system and medium based on big data. BACKGROUND
[0002] With the rapid development of metaverse technology, the fusion of virtual scenes and the real world is increasingly close. Metaverse e-commerce (such as virtual toy blind boxes and digital collectible transactions) that takes "immersive interaction and virtual-real combination" as the core has become a new business form. User behavior presents multi-dimensional, high-dynamic and cross-device characteristics. Behavior data sources are complex, user interaction involves multiple channels such as metaverse clients, VR / AR hardware devices, blockchain digital wallets and virtual social systems, and there are large differences in data formats and high dispersion. Traditional analysis techniques are not adaptable, and existing user behavior analysis methods mainly target offline physical scenes or traditional 2D e-commerce, such as click, stay and order data, and do not cover core data such as virtual space movement trajectories, immersive interaction duration and virtual identity-related behaviors specific to metaverse, resulting in single analysis dimension and low scene fit. Data integration and value conversion fault, metaverse multi-source data has problems such as device timestamp deviation, difficulty in virtual coordinate system unification and non-uniform identity identification, and existing data processing techniques cannot eliminate data differences. At the same time, the linkage between analysis results and metaverse scene optimization, such as virtual booth layout and VR opening box interaction design, and operation strategies, such as digital collectible pushing, is weak, and a complete closed loop of data collection, analysis and application cannot be formed. Therefore, there is an urgent need for a user behavior analysis method that adapts to the characteristics of metaverse scenes and can realize multi-source data integration and full-process value conversion.
[0003] In view of the above problems, an effective technical solution is currently needed. SUMMARY
[0004] The purpose of the present application is to provide a metaverse user behavior analysis closed loop method, system and medium based on big data, which can realize closed loop analysis of metaverse user behavior through accurate data acquisition and efficient integration, behavior rule mining, dynamic monitoring and closed loop iterative optimization that adapt to the scene.
[0005] In a first aspect, the present application provides a metaverse user behavior analysis closed loop method based on big data, comprising the following steps: Collecting multi-modal data of metaverse users to obtain a multi-modal data set, including virtual space operation record data, digital collectible interaction record data and user association data; acquire preset scene data preprocessing standard rules, preprocess the virtual space operation record data, digital collectible interaction record data and user association data according to the preset scene data preprocessing standard rules, and obtain virtual space operation record standard data, digital collectible interaction record standard data and user association standard data; perform data bit rectification processing on the virtual space operation record standard data, digital collectible interaction record standard data and user association standard data, and generate user behavior archive data; According to the user behavior archive data, a funnel model is used for disassembly and user traceability sorting processing, and a business process funnel report and a user grouping and value analysis table are generated; According to the user behavior archive data, the business process funnel report and the user grouping and value analysis table are analyzed and processed, and real-time index monitoring report and abnormal log record data are obtained; Input the real-time index monitoring report and the abnormal log record data into a preset behavior prediction model for processing, and obtain user individual behavior prediction data and group user prediction trend data; According to the user individual behavior prediction data and the group user prediction trend data, the initial business scheme is optimized, and an optimized business scheme is obtained.
[0006] Optionally, in the meta-universe user behavior analysis closed loop method based on big data provided in the present application, the multi-modal data of the meta-universe user is collected to obtain a multi-modal data set, including virtual space operation record data, digital collectible interaction record data and user association data, including: Collect multi-modal data of the meta-universe user to obtain a multi-modal data set, including virtual space operation record data, digital collectible interaction record data and user association data; The virtual space operation record data includes movement trajectory, virtual exhibition stand coordinates, opening box gesture stay time and opening box force; The digital collectible interaction record data includes digital blind box ID, opening box operation type feature data, opening box frequency, collectible consumption amount, collectible viewing time record data, sharing times and virtual interaction rate; The user association data includes user virtual identity ID, user level and historical opening box IP preference.
[0007] Optionally, in the meta-universe user behavior analysis closed loop method based on big data provided in the present application, the preset scene data preprocessing standard rules are acquired, and the virtual space operation record data, digital collectible interaction record data and user association data are preprocessed according to the preset scene data preprocessing standard rules. The virtual space operation record standard data, the digital collectible interaction record standard data and the user association standard data are obtained, including: Obtaining preset scenario data preprocessing standard rules, including data format unification rules, data cleaning rules, field definition unification rules, data correlation rules and business adaptation rules; According to the preset scenario data preprocessing standard rules, the virtual space operation record data, the digital collection interaction record data and the user association data are timestamped, converted and preprocessed, and virtual coordinate conversion and abnormal data cleaning are obtained. Virtual space operation record standard data, digital collection interaction record standard data and user association standard data.
[0008] Optionally, in the big data-based meta-universe user behavior analysis closed loop method described in the present application, the virtual space operation record standard data, the digital collection interaction record standard data and the user association standard data are subjected to data bit rectification processing to generate user behavior profile data, comprising: According to the user virtual identity ID, the virtual space operation record standard data, the digital collection interaction record standard data and the user association standard data are correlated to generate initial user behavior profile data; The initial user behavior profile data is subjected to device timestamp deviation correction and virtual coordinate unification data bit rectification processing to generate user behavior profile data.
[0009] Optionally, in the big data-based meta-universe user behavior analysis closed loop method described in the present application, the user behavior profile data is subjected to funnel model disassembly and user traceability group sorting processing to generate a business process funnel report and a user group and value analysis table, comprising: According to the user behavior profile data, a funnel model disassembly analysis is performed to obtain a conversion rate of the link; The conversion rate of the link is compared with a preset link standard conversion rate; If the conversion rate of the link is less than the preset link standard conversion rate, it is determined as a user loss node; If the conversion rate of the link is greater than or equal to the preset link standard conversion rate, it is determined as a normal node; According to the conversion rate of the link and the user loss node or the normal node, a business process funnel report is generated; According to the user virtual identity ID, the registration duration of the user is obtained; The user level, registration duration, box opening frequency and historical box opening IP preference are input into a preset user group analysis model for processing to obtain a user group, including new users, potential loss users, stable active users, high-value core users or silent low-value users; The collection consumption amount and virtual interaction rate corresponding to the user group are weighted and summed to obtain a comprehensive value evaluation index corresponding to the user group, and are arranged in descending order to generate a user group value sorting table. generating a user cluster and value analysis table according to the user cluster and user cluster value ranking table.
[0010] Optionally, in the big data-based meta-universe user behavior analysis closed-loop method described in the present application, the analysis and processing according to the user behavior archive data in combination with the business process funnel report and the user cluster and value analysis table obtains real-time index monitoring reports and abnormal log record data, including: According to the user behavior archive data, data extraction is performed to obtain virtual blind box opening conversion rate, digital collectible repurchase rate, opening box process abandonment rate, continuous non-entry scene days, virtual exhibition stand stay duration, and cooperative opening frequency corresponding to the user cluster; According to the virtual blind box opening conversion rate, digital collectible repurchase rate, opening box process abandonment rate, continuous non-entry scene days, virtual exhibition stand stay duration, and cooperative opening frequency, a real-time index monitoring report is generated; According to the business process funnel report, the initial baseline threshold value corresponding to the virtual blind box opening conversion rate, digital collectible repurchase rate, opening box process abandonment rate, continuous non-entry scene days, virtual exhibition stand stay duration, and cooperative opening frequency is extracted; According to the user cluster and value analysis table, a threshold correction coefficient is determined, and the initial baseline threshold value is corrected to obtain the baseline threshold value corresponding to the virtual blind box opening conversion rate, digital collectible repurchase rate, opening box process abandonment rate, continuous non-entry scene days, virtual exhibition stand stay duration, and cooperative opening frequency; The virtual blind box opening conversion rate, digital collectible repurchase rate, opening box process abandonment rate, continuous non-entry scene days, virtual exhibition stand stay duration, or cooperative opening frequency is compared with the corresponding baseline threshold value, and abnormal log record data is generated according to the threshold comparison result.
[0011] Optionally, in the big data-based meta-universe user behavior analysis closed-loop method described in the present application, the real-time index monitoring report and the abnormal log record data are input into a preset behavior prediction model for processing to obtain user individual behavior prediction data and group user prediction trend data, including: According to the user behavior archive data, data extraction is performed to obtain historical behavior feature data, including opening frequency mean and collectible repurchase interval mean; The opening frequency mean and collectible repurchase interval mean, as well as the real-time index monitoring report and the abnormal log record data, are input into a preset behavior prediction model for processing to obtain user individual behavior prediction data and group user prediction trend data.
[0012] Optionally, in the big data-based meta-universe user behavior analysis closed-loop method described in the present application, the step of optimizing the initial business scheme according to the user individual behavior prediction data and the group user prediction trend data to obtain an optimized business scheme comprises: optimizing the initial business scheme according to the user individual behavior prediction data and the group user prediction trend data to obtain an optimized business scheme; The optimized business scheme comprises virtual scene optimization, operation strategy optimization and product function optimization.
[0013] In a second aspect, the present application provides a big data-based meta-universe user behavior analysis closed-loop system, which comprises a memory and a processor, wherein the memory comprises a program of a big data-based meta-universe user behavior analysis closed-loop method, and the program of the big data-based meta-universe user behavior analysis closed-loop method is executed by the processor to realize the following steps: collecting multi-modal data of meta-universe users to obtain a multi-modal data set, comprising virtual space operation record data, digital collectible interaction record data and user association data; obtaining preset scene data preprocessing standard rules, and preprocessing the virtual space operation record data, the digital collectible interaction record data and the user association data according to the preset scene data preprocessing standard rules to obtain virtual space operation record standard data, digital collectible interaction record standard data and user association standard data; performing data bit rectification processing on the virtual space operation record standard data, the digital collectible interaction record standard data and the user association standard data to generate user behavior profile data; generating a business process funnel report and a user grouping and value analysis table through funnel model disassembly and user traceability grouping sorting processing according to the user behavior profile data; analyzing and processing the user behavior profile data in combination with the business process funnel report and the user grouping and value analysis table to obtain real-time index monitoring reports and abnormal log record data; inputting the real-time index monitoring reports and the abnormal log record data into a preset behavior prediction model for processing to obtain user individual behavior prediction data and group user prediction trend data; optimizing the initial business scheme according to the user individual behavior prediction data and the group user prediction trend data to obtain an optimized business scheme.
[0014] In a third aspect, the present application further provides a computer readable storage medium, wherein a big data-based meta-universe user behavior analysis closed-loop method program is stored in the computer readable storage medium, and the big data-based meta-universe user behavior analysis closed-loop method program is executed by a processor to realize the steps of the big data-based meta-universe user behavior analysis closed-loop method according to any one of the above aspects.
[0015] From the above, the big data-based meta-universe user behavior analysis closed loop method, system and medium provided by the application realize closed loop analysis of meta-universe user behavior through accurate data acquisition and efficient integration of adaptive scenes, behavior rule mining, dynamic monitoring and closed loop iterative optimization.
[0016] Other features and advantages of the present application will be described in the following description, and some will become apparent from the description, or will be understood by those skilled in the art. The purpose and other advantages of the present application can be achieved and obtained by the structure specifically pointed out in the written description and drawings. BRIEF DESCRIPTION OF DRAWINGS
[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as limiting the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.
[0018] Figure 1 The flowchart of the big data-based meta-universe user behavior analysis closed loop method provided by the embodiments of the present application is shown in the following figure: Figure 2 The flowchart of generating user behavior archive data of the big data-based meta-universe user behavior analysis closed loop method provided by the embodiments of the present application is shown in the following figure: Figure 3 The flowchart of generating business process funnel report and user grouping and value analysis table of the big data-based meta-universe user behavior analysis closed loop method provided by the embodiments of the present application is shown in the following figure: Figure 4 The high-level flowchart of the big data-based meta-universe user behavior analysis closed loop method provided by the embodiments of the present application is shown in the following figure. DETAILED DESCRIPTION
[0019] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.
[0020] It should be noted that similar reference numbers and letters refer to similar items in the following drawings, and therefore, once an item is defined in one drawing, it need not be further defined and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms "first", "second", and the like are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.
[0021] Please refer to Figure 1 , Figure 1 is a flowchart of a big data-based meta-universe user behavior analysis closed-loop method in some embodiments of the present application. The big data-based meta-universe user behavior analysis closed-loop method is used in a terminal device, such as a computer, a mobile phone terminal, etc. The big data-based meta-universe user behavior analysis closed-loop method includes the following steps: S11, collecting multi-modal data of meta-universe users to obtain a multi-modal data set, including virtual space operation record data, digital collectible interaction record data, and user association data; S12, obtaining a preset scenario data preprocessing standard rule, preprocessing the virtual space operation record data, the digital collectible interaction record data, and the user association data according to the preset scenario data preprocessing standard rule, to obtain virtual space operation record standard data, digital collectible interaction record standard data, and user association standard data; S13, performing data bit rectification processing on the virtual space operation record standard data, the digital collectible interaction record standard data, and the user association standard data to generate user behavior profile data; S14, according to the user behavior profile data, processing through a funnel model disassembly and user traceability group sorting, to generate a business process funnel report and a user group and value analysis table; S15, according to the user behavior profile data, combining the business process funnel report and the user group and value analysis table for analysis processing, to obtain a real-time index monitoring report and abnormal log record data; S16, inputting the real-time index monitoring report and the abnormal log record data into a preset behavior prediction model for processing, to obtain user individual behavior prediction data and group user prediction trend data; S17, optimizing an initial business scheme according to the user individual behavior prediction data and the group user prediction trend data, to obtain an optimized business scheme.
[0022] It should be noted that in order to accurately construct the user behavior analysis method of the Yuanqi Matte metaverse scene, first, a multi-modal data set of game player related virtual space operation record data, digital collectible interaction record data and user associated data is collected, then the corresponding standard data is obtained according to the matched preset scene data preprocessing standard rule, and the data bit is corrected, a unified user behavior profile data is generated, the business process is disassembled through the funnel model to mine user behavior rules, analyze the conversion rate of each link, generate a business process funnel report, and at the same time, the user is divided into groups and the value is analyzed to generate a user group and value analysis table, and then taking the business process funnel report and the user group and value analysis table as the benchmark, the core index monitoring and analysis are carried out, the real-time index monitoring report and the abnormal log record data are obtained, the pre-trained preset behavior prediction model is combined for processing, the user individual behavior prediction data and the group user prediction trend data are obtained, which are used to evaluate individual behavior and group trend, and finally the scheme is optimized according to the analysis, forming a complete iterative closed loop of data collection and processing, analysis, prediction and optimization.
[0023] According to the embodiment of the application, the multi-modal data of the metaverse user is collected to obtain a multi-modal data set, including virtual space operation record data, digital collectible interaction record data and user associated data, which includes: The multi-modal data of the metaverse user is collected to obtain a multi-modal data set, including virtual space operation record data, digital collectible interaction record data and user associated data; The virtual space operation record data includes movement trajectory, virtual exhibition stand coordinates, opening box gesture stay time and opening box force; The digital collectible interaction record data includes digital blind box ID, opening box operation type feature data, opening box frequency, collectible consumption amount, collectible viewing time record data, sharing times and virtual interaction rate; The user associated data includes user virtual identity ID, user level and historical opening box IP preference.
[0024] It should be noted that the multi-modal data of different users in the virtual scene in a preset time period (such as the past 30 days) is collected through the data collection module, wherein the movement trajectory refers to the position transformation path of the meta-universe user in the virtual scene, such as the path from the store entrance to the AIP blind box exhibition stand, the virtual exhibition stand coordinate refers to the stay position coordinate of the user at the virtual exhibition stand, the opening gesture stay duration refers to the total duration from the start of the opening gesture to the completion of the opening, the opening force refers to the sensed hand force value, the opening operation type refers to the VR gesture opening, the client click opening or the collaborative opening, the opening operation type feature data is represented by different values, the opening frequency refers to the number of openings in a preset time period (such as one day), the virtual interaction rate refers to the ratio of the daily interaction frequency to the online duration, and the historical opening IP preference is evaluated by the interaction data of the user in the virtual scene to evaluate the user's preference for different IP series, such as original virtual image IP, cross-border joint IP, designer joint IP or scene theme IP.
[0025] According to the embodiment of the application, the preset scene data preprocessing standard rule is obtained, the virtual space operation record data, the digital collectible interaction record data and the user association data are preprocessed according to the preset scene data preprocessing standard rule, and virtual space operation record standard data, digital collectible interaction record standard data and user association standard data are obtained, including: The preset scene data preprocessing standard rule includes data format unification rule, data cleaning rule, field definition unification rule, data association rule and business adaptation rule; The virtual space operation record data, the digital collectible interaction record data and the user association data are preprocessed by timestamp unification, virtual coordinate conversion and abnormal data cleaning according to the preset scene data preprocessing standard rule, and virtual space operation record standard data, digital collectible interaction record standard data and user association standard data are obtained.
[0026] It should be noted that the data format unification rule is used to ensure that the formats of data from different sources are consistent, such as VR devices, clients and blockchains, the data cleaning rule is used to eliminate worthless abnormal data and fill in missing data, the field definition unification rule is used to standardize the naming and meaning of the core behavior fields of the meta-universe toy scene, to ensure that different technical personnel have consistent understanding of the data, the data association rule binds the data scattered in different systems through the user virtual identity ID, to form a complete user behavior link file, and the business adaptation rule formulates exclusive rules for the special business scenes of the meta-universe toy, to ensure that the data processing result can directly support business decision-making, such as collaborative opening. The obtained multi-modal data set is preprocessed according to the determined preset scene data preprocessing standard rule, and corresponding standard data is obtained, including virtual space operation record standard data, digital collectible interaction record standard data and user association standard data.
[0027] Please refer to Figure 2 , Figure 2 is a flowchart of generating user behavior archive data in the big data-based meta-universe user behavior analysis closed-loop method in some embodiments of the present application. According to the embodiment of the present application, the virtual space operation record standard data, digital collection interaction record standard data and user association standard data are subjected to data bit rectification processing to generate user behavior archive data, which comprises: S21, according to the user virtual identity ID, the virtual space operation record standard data, the digital collection interaction record standard data and the user association standard data are associated to generate initial user behavior archive data; S22, the initial user behavior archive data is subjected to device timestamp deviation correction and virtual coordinate unification data bit rectification processing to generate user behavior archive data.
[0028] It should be noted that, first, the preprocessed standard data, including the standard data recording the user's operation record in the virtual space, the digital collection interaction record standard data reflecting the core operation of the collection, and the user association standard data containing the basic attributes of the user are directly and accurately associated according to the user virtual identity ID to generate initial user behavior archive data containing user space operation, collection interaction and basic attributes, to solve the problem of multi-modal data set fragmentation, then, calibrate multi-device time to UTC+8 millisecond level timestamp, at the same time, map different types of coordinates to the plane rectangular coordinate system of the virtual toy store of the meta universe, realize data bit rectification processing, realize the rectification of inconsistent data, finally, generate precise and logically coherent user behavior archive data which can be directly used for subsequent funnel report and user grouping and value analysis.
[0029] Please refer to Figure 3 , Figure 3 is a flowchart of generating business process funnel report and user grouping and value analysis table in the big data-based meta-universe user behavior analysis closed-loop method in some embodiments of the present application. According to the embodiment of the present application, the user behavior archive data is subjected to funnel model disassembly and user traceability grouping and sorting processing to generate business process funnel report and user grouping and value analysis table, which comprises: S31, according to the user behavior archive data, the funnel model disassembly analysis processing is performed to obtain the link conversion rate; S32, the link conversion rate is compared with the preset link standard conversion rate; S321, if the link conversion rate is less than the preset link standard conversion rate, it is determined as a user loss node; S322, if the link conversion rate is greater than or equal to the preset link standard conversion rate, it is determined as a normal node; S33, generating a business process funnel report according to the link conversion rate and the user churn node or normal node; S34, obtaining the registration duration of the user according to the user virtual identity ID; S35, inputting the user level, registration duration, unpacking frequency and historical unpacking IP preference into a preset user clustering analysis model for processing to obtain user clustering, including new users, potential churn users, stable active users, high-value core users or silent low-value users; S36, performing weighted summation on the collection consumption amount and virtual interaction rate corresponding to the user clustering to obtain a comprehensive value evaluation index corresponding to the user clustering, and performing descending arrangement to generate a user clustering value sorting table; S37, generating a user clustering and value analysis table according to the user clustering and the user clustering value sorting table.
[0030] It should be noted that by dividing the core business process of the metaverse scene into links and analyzing the conversion rate of each link, the churn node can be accurately determined to provide basic data for subsequent business optimization. The link conversion rate is the ratio of the number of conversions in the current link to the number of entries in the previous link. For example, 50 people enter a virtual store, 10 people choose a blind box exhibition stand, and 5 people trigger unpacking interaction. Then 5 / 10=0.5 is the link conversion rate. According to the link conversion rate of different links, the normal node or the user churn node is determined by threshold comparison, and a business process funnel report is generated accordingly. At the same time, different users are divided into groups by a preset user clustering analysis model according to the user level, registration duration, unpacking frequency and historical unpacking IP preference, and the corresponding collection consumption amount and virtual interaction rate are weighted and summed to sort the value and generate a user clustering and value analysis table. The preset user clustering analysis model is trained by obtaining a large number of historical sample user levels, registration durations, unpacking frequencies and historical unpacking IP preferences and corresponding user clustering.
[0031] According to the embodiments of the present application, the user behavior profile data is analyzed and processed according to the business process funnel report and the user clustering and value analysis table to obtain real-time index monitoring reports and abnormal log record data, including: According to the user behavior profile data, data extraction is performed to obtain the virtual blind box unpacking conversion rate, digital collection repurchase rate, unpacking process abandonment rate, consecutive non-scene entry days, virtual exhibition stand stay duration and collaborative unpacking frequency corresponding to the user clustering; According to the virtual blind box unpacking conversion rate, digital collection repurchase rate, unpacking process abandonment rate, consecutive non-scene entry days, virtual exhibition stand stay duration and collaborative unpacking frequency, real-time index monitoring reports are generated; According to the business process funnel report, the initial baseline threshold values corresponding to the virtual blind box opening conversion rate, the digital collectible repurchase rate, the opening box process abandonment rate, the continuous non-entry scene day number, the virtual exhibition stand stay duration, and the cooperative opening frequency are extracted; According to the user group and value analysis table, the threshold correction coefficient is determined, and the initial baseline threshold value is corrected to obtain the baseline threshold value corresponding to the virtual blind box opening conversion rate, the digital collectible repurchase rate, the opening box process abandonment rate, the continuous non-entry scene day number, the virtual exhibition stand stay duration, and the cooperative opening frequency; The virtual blind box opening conversion rate, the digital collectible repurchase rate, the opening box process abandonment rate, the continuous non-entry scene day number, the virtual exhibition stand stay duration, or the cooperative opening frequency is compared with the corresponding baseline threshold value, and the threshold comparison result is used to generate abnormal log record data.
[0032] It should be noted that, first, the core indicators are determined from the three aspects of conversion, loss and interaction, and real-time monitoring is carried out according to different user groups, the virtual blind box opening conversion rate, the digital collection repurchase rate, the opening box process abandonment rate, the continuous non-entry scene days, the virtual exhibition stand stay time and the cooperative opening frequency are the mean values of the corresponding user groups, wherein the virtual blind box opening conversion rate is the ratio of the actual opening box interaction user number to the selected blind box user number in a preset time period (such as 15 days), the digital collection repurchase rate is the ratio of the users who purchase the same type of digital collection again to the total users in a preset time period (such as 30 days) after purchasing a certain type of digital collection, the opening box process abandonment rate is the ratio of the number of users who abandon before completing the opening box viewing operation to the number of users who trigger the blind box opening, and the cooperative opening frequency is the ratio of the number of users who initiate opening boxes with friends to the total opening box users in a preset time period (such as 30 days), then, the corresponding initial reference threshold is queried, the initial reference threshold is the mean value corresponding to the user group multiplied by a preset business target adjustment coefficient (determined by the person skilled in the art according to historical sample analysis), and can be dynamically adjusted, at the same time, the threshold correction coefficient is determined according to the user group and the value analysis table, the threshold correction coefficient is dynamically set by the person skilled in the art according to the core indicators and the user group, for example, the digital collection repurchase rate of high-value core users is 55%, the digital collection repurchase rate of new users is 10%, the preset business target adjustment coefficient is 1.03, the threshold correction coefficient of high-value core users is set to 1.1, and the corresponding reference threshold is 55% x 1.03 x 1.1 = 62.3%, the purpose is to retain high-value users, and the threshold correction coefficient of new users is set to 1.2, the purpose is to set a higher improvement coefficient according to the small base of the digital collection repurchase rate of new users, and the initial reference threshold is corrected to determine the corresponding reference threshold, the virtual blind box opening conversion rate, the digital collection repurchase rate, the opening box process abandonment rate, the continuous non-entry scene days, the virtual exhibition stand stay time or the cooperative opening frequency are compared with the corresponding reference threshold, and the threshold comparison is not met (the virtual blind box opening conversion rate, the digital collection repurchase rate, the virtual exhibition stand stay time and the cooperative opening frequency are less than the threshold, and the opening box process abandonment rate and the continuous non-entry scene days are greater than the threshold). Record and generate abnormal log record data.
[0033] According to the embodiment of the application, the real-time index monitoring report and the abnormal log record data are input into a preset behavior prediction model for processing to obtain user individual behavior prediction data and group user prediction trend data, which comprises: According to the user behavior profile data, historical behavior feature data including opening box frequency mean value and collection repurchase interval mean value are obtained; The box opening frequency mean value and the collection repeat purchase interval mean value, and the real-time index monitoring report and the abnormal log record data are input into a preset behavior prediction model for processing to obtain user individual behavior prediction data and group user prediction trend data.
[0034] It should be noted that, according to the box opening frequency mean value and the collection repeat purchase interval mean value of the reaction user historical behavior rule, the real-time index monitoring report and the abnormal log record data of the real-time behavior and the sudden factor are analyzed by the preset behavior prediction model to obtain user individual behavior prediction data (for example, the user A opening box probability is 80% in the future 7 days) focusing on the specific behavior of a certain user and group user prediction trend data (for example, in the future 7 days, the high-value core user group, the repeat purchase rate of the new online CIP series blind box is 60%) focusing on the user group, wherein the preset behavior prediction model is trained by inputting a large number of historical sample box opening frequency mean value and collection repeat purchase interval mean value, real-time index monitoring report and abnormal log record data and corresponding user individual behavior prediction data and group user prediction trend data into an LSTM neural network model under a TensorFlow framework.
[0035] According to the embodiment of the present application, the initial business scheme is optimized according to the user individual behavior prediction data and the group user prediction trend data to obtain an optimized business scheme, which comprises: According to the user individual behavior prediction data and the group user prediction trend data, the initial business scheme is optimized to obtain an optimized business scheme. The optimized business scheme comprises virtual scene optimization, operation strategy optimization and product function optimization.
[0036] It should be noted that according to the prediction result, the business scheme is realized to optimize the virtual scene, the operation strategy and the product function. The main purpose of the virtual scene optimization is to improve the immersive interaction experience of the user and reduce the conversion anomaly. For example, if the group user prediction trend data shows that the longest stay time is at the BIP series exhibition stand but the opening box conversion rate is low, and the user individual behavior prediction data indicates that 65% of the high-value core users give up opening the box due to the crowded interaction area of the exhibition stand, the virtual scene optimization scheme can be adjusted to adjust the visual layout of the BIP series virtual exhibition stand from 4 square centimeters to 9 square centimeters, and improve the light effect to enhance the attraction to the user. The main purpose of the operation strategy optimization is to stimulate the consumption and interaction willingness according to the user grouping. For example, according to the user X's opening box probability of 95% in the next 3 days and the preference for AIP hidden type, the corresponding operation strategy is optimized, and the user D is pushed to purchase the AIP hidden type blind box priority and appropriate discount, and the opening frequency and sharing conversion rate of the user are improved through the group interaction amplification effect. The purpose of the product function optimization is to solve the abnormal point and improve the practicability. For example, the user individual behavior prediction data shows that 35% of the users give up opening the box due to the waiting process because the VR opening box animation loading time is more than 6 seconds, and the abnormal log record frequency is high. Therefore, the product function is optimized, such as compressing the VR opening box animation file size, and using the preloading technology to shorten the loading time from 6 seconds to 2 seconds, thereby improving the virtual use experience and function stickiness of the user.
[0037] Please refer to Figure 4 , Figure 4 is a high-level flowchart of a big data-based meta-universe user behavior analysis closed-loop method in some embodiments of the present application.
[0038] The application also discloses a big data-based meta-universe user behavior analysis closed-loop system, comprising a memory and a processor, wherein the memory comprises a big data-based meta-universe user behavior analysis closed-loop method program, and the big data-based meta-universe user behavior analysis closed-loop method program is executed by the processor to realize the following steps: Collecting multi-modal data of meta-universe users to obtain a multi-modal data set, comprising virtual space operation record data, digital collectible interaction record data and user association data; Obtaining a preset scene data preprocessing standard rule, preprocessing the virtual space operation record data, the digital collectible interaction record data and the user association data according to the preset scene data preprocessing standard rule to obtain virtual space operation record standard data, digital collectible interaction record standard data and user association standard data; Performing data bit correction processing on the virtual space operation record standard data, the digital collectible interaction record standard data and the user association standard data to generate user behavior profile data; According to the user behavior archive data, a funnel model is used to analyze and sort the user traceability groups, a business process funnel report and a user group and value analysis table are generated; According to the user behavior archive data, a funnel model is used to analyze and sort the user traceability groups, a business process funnel report and a user group and value analysis table are generated; According to the user behavior archive data, a funnel model is used to analyze and sort the user traceability groups, a business process funnel report and a user group and value analysis table are generated; According to the user behavior archive data, a funnel model is used to analyze and sort the user traceability groups, a business process funnel report and a user group and value analysis table are generated.
[0039] It should be noted that, in order to accurately construct the user behavior analysis method of the Meta universe scene, first, a multi-modal data set of game player virtual space operation record data, digital collectible interaction record data and user association data is collected, then the corresponding standard data is obtained according to the matched preset scene data preprocessing standard rule, and the data bit is corrected to generate unified user behavior archive data, the business process is disassembled to mine user behavior rules through a funnel model, the conversion rate of each link is analyzed, a business process funnel report is generated, and the user is divided into groups and the value is analyzed to generate a user group and value analysis table. According to the business process funnel report and the user group and value analysis table, the core indicators are monitored and analyzed to obtain real-time indicator monitoring report and abnormal log record data, and the pre-trained preset behavior prediction model is processed to obtain user individual behavior prediction data and group user prediction trend data, which is used to evaluate individual behavior and group trend, and finally the scheme is optimized according to the analysis to form a complete iterative closed loop of data collection and processing, analysis, prediction and optimization.
[0040] According to the embodiment of the application, the multi-modal data of the Meta universe user is collected to obtain a multi-modal data set, including virtual space operation record data, digital collectible interaction record data and user association data, including: The multi-modal data of the Meta universe user is collected to obtain a multi-modal data set, including virtual space operation record data, digital collectible interaction record data and user association data; The virtual space operation record data includes movement trajectory, virtual exhibition stand coordinates, opening box gesture stay time and opening box force; The digital collectible interaction record data includes digital blind box ID, opening box operation type feature data, opening box frequency, collectible consumption amount, collectible viewing time record data, sharing times and virtual interaction rate; The user association data includes user virtual identity ID, user level and historical opening box IP preference.
[0041] It should be noted that the multi-modal data of different users in the virtual scene in a preset time period (such as the past 30 days) is collected through the data collection module, wherein the movement trajectory refers to the position transformation path of the meta-universe user in the virtual scene, such as the path from the store entrance to the AIP blind box exhibition stand, the virtual exhibition stand coordinate refers to the stay position coordinate of the user at the virtual exhibition stand, the opening gesture stay duration refers to the total duration from the start of the opening gesture to the completion of the opening, the opening force refers to the sensed hand force value, the opening operation type refers to the VR gesture opening, the client click opening or the collaborative opening, the opening operation type feature data is represented by different values, the opening frequency refers to the number of openings in a preset time period (such as one day), the virtual interaction rate refers to the ratio of the average daily interaction times to the online duration, and the historical opening IP preference is evaluated by the user's interaction data in the virtual scene, such as the user's preference for different IP series, such as original virtual image IP, cross-border joint IP, designer joint IP or scene theme IP.
[0042] According to the embodiment of the application, the preset scene data preprocessing standard rule is obtained, the virtual space operation record data, the digital collectible interaction record data and the user association data are preprocessed according to the preset scene data preprocessing standard rule, and the virtual space operation record standard data, the digital collectible interaction record standard data and the user association standard data are obtained, including: The preset scene data preprocessing standard rule includes data format unification rule, data cleaning rule, field definition unification rule, data association rule and business adaptation rule. According to the preset scene data preprocessing standard rule, the virtual space operation record data, the digital collectible interaction record data and the user association data are preprocessed by time stamp unification, virtual coordinate conversion and abnormal data cleaning, and the virtual space operation record standard data, the digital collectible interaction record standard data and the user association standard data are obtained.
[0043] It should be noted that the unified data format rules are used to ensure the consistency of the format of data from different sources, such as VR devices, clients and blockchains. The data cleaning rules are used to eliminate worthless abnormal data and fill in missing data. The unified field definition rules are used to standardize the naming and meaning definition of the core behavior fields of the metaverse trendy play scene to ensure that different technical personnel have a consistent understanding of the data. The data association rules bind data scattered in different systems through the user's virtual identity ID to form a complete user behavior link archive. The business adaptation rules formulate exclusive rules for the unique business scenarios of metaverse trendy plays to ensure that the data processing results can directly support business decisions. For example, for collaborative unpacking, the acquired multimodal data set is preprocessed according to the determined preset scenario-based data preprocessing standard rules to obtain the corresponding standard data, including virtual space operation record standard data, digital collection interaction record standard data and user association standard data.
[0044] According to an embodiment of the present invention, performing data bit deviation correction processing on the virtual space operation record standard data, the digital collection interaction record standard data, and the user association standard data to generate user behavior profile data includes: According to the user's virtual identity ID, the virtual space operation record standard data, the digital collection interaction record standard data and the user association standard data are associated with each other to generate initial user behavior profile data; The initial user behavior file data is subjected to a data bit deviation correction process for device timestamp deviation correction and virtual coordinate unification to generate user behavior file data.
[0045] It should be noted that, first of all, the pre-processed standard data, including the standard data recording the user's operation records in the virtual space, the standard data reflecting the digital collection interaction records of the core operations of the collection, and the user association standard data containing the user's basic attributes are directly and accurately associated according to the user's virtual identity ID to generate the initial user behavior profile data containing user space operations, collection interactions and basic attributes to solve the problem of fragmentation of multimodal data sets. Then, the multi-device time is calibrated to a unified UTC+8 millisecond timestamp. At the same time, the coordinates of different categories are mapped to the plane rectangular coordinate system of the Yuanqi Mart virtual trendy toy store to realize data bit correction processing and correct the inconsistent data. Finally, accurate and logically coherent user behavior profile data is generated that can be directly used for subsequent funnel reports and user segmentation and value analysis.
[0046] According to an embodiment of the present invention, the user behavior profile data is decomposed through a funnel model and sorted by user tracing and grouping to generate a business process funnel report and a user grouping and value analysis table, including: According to the user behavior profile data, the conversion rate of each link is obtained by disassembling and analyzing the data through the funnel model; threshold comparison is performed between the link conversion rate and a preset link standard conversion rate; If the link conversion rate is less than the preset link standard conversion rate, it is determined as a user loss node; If the link conversion rate is greater than or equal to the preset link standard conversion rate, it is determined as a normal node; According to the link conversion rate and the user loss node or normal node, a business process funnel report is generated; According to the user virtual identity ID, the registration duration of the user is obtained; The user level, registration duration, box opening frequency and historical box opening IP preference are input into a preset user clustering analysis model for processing to obtain user clusters, including new users, potential loss users, stable active users, high-value core users or silent low-value users; The collection consumption amount and virtual interaction rate corresponding to the user clusters are weighted and summed to obtain a comprehensive value evaluation index corresponding to the user clusters, and the comprehensive value evaluation index is arranged in descending order to generate a user cluster value sorting table; According to the user cluster and the user cluster value sorting table, a user cluster and value analysis table is generated.
[0047] It should be noted that by dividing the core business process of the meta universe scene into links and analyzing the conversion rate of each link, the loss node can be accurately determined to provide basic data for subsequent business optimization. The link conversion rate refers to the ratio of the number of conversions in the current link to the number of people entering the previous link. For example, 50 people enter a virtual store, 10 people choose a blind box exhibition stand, and 5 people trigger the box opening interaction. Then, 5 / 10=0.5 is the link conversion rate. According to the different links, the link conversion rate is calculated, and then the threshold comparison is used to determine the normal node or the user loss node, and a business process funnel report is generated accordingly. At the same time, different users are divided into groups according to the user level, registration duration, box opening frequency and historical box opening IP preference through a preset user clustering analysis model, and the corresponding collection consumption amount and virtual interaction rate are weighted and summed to sort the value, and a user cluster and value analysis table is generated. The preset user clustering analysis model is trained by obtaining a large number of historical sample user levels, registration durations, box opening frequencies and historical box opening IP preferences and corresponding user clusters.
[0048] According to the embodiment of the application, the user behavior profile data is analyzed and processed in combination with the business process funnel report and the user cluster and value analysis table to obtain real-time index monitoring reports and abnormal log record data, including: According to the user behavior profile data, data extraction is performed to obtain virtual blind box opening conversion rate, digital collectible repurchase rate, opening box process abandonment rate, continuous non-entry scene day number, virtual exhibition stand stay duration, and cooperative opening frequency corresponding to the user cluster; According to the virtual blind box opening conversion rate, digital collectible repurchase rate, opening box process abandonment rate, continuous non-entry scene day number, virtual exhibition stand stay duration, and cooperative opening frequency, a real-time index monitoring report is generated; According to the business process funnel report, an initial baseline threshold value corresponding to the virtual blind box opening conversion rate, digital collectible repurchase rate, opening box process abandonment rate, continuous non-entry scene day number, virtual exhibition stand stay duration, and cooperative opening frequency is extracted; According to the user cluster and value analysis table, a threshold correction coefficient is determined, and the initial baseline threshold value is corrected to obtain a baseline threshold value corresponding to the virtual blind box opening conversion rate, digital collectible repurchase rate, opening box process abandonment rate, continuous non-entry scene day number, virtual exhibition stand stay duration, and cooperative opening frequency; The virtual blind box opening conversion rate, digital collectible repurchase rate, opening box process abandonment rate, continuous non-entry scene day number, virtual exhibition stand stay duration, or cooperative opening frequency is compared with the corresponding baseline threshold value, and an abnormal log record data is generated according to the threshold comparison result.
[0049] It should be noted that, first of all, the core indicators are determined from the three aspects of conversion, churn and interaction, and real-time monitoring is carried out according to different user groups. The virtual blind box opening conversion rate, digital collection repurchase rate, unboxing process abandonment rate, consecutive days of not entering the scene, virtual booth stay time and collaborative unboxing frequency are all the average values of the corresponding user groups. Among them, the virtual blind box opening conversion rate refers to the ratio of the number of actual unboxing interaction users within a preset time period (such as 15 days) to the number of users who selected the blind box. The digital collection repurchase rate refers to the number of users who purchase a certain type of digital collection again within a preset time period (such as 30 days). The ratio of users of the type of digital collections to the total number of users; the abandonment rate of the unboxing process refers to the ratio of the number of users who give up before triggering the blind box opening to completing the unboxing and viewing operation to the number of users who triggered the blind box opening; the frequency of collaborative unboxing refers to the ratio of the number of users who jointly initiate unboxing with friends within a preset time period (such as 30 days) to the total number of unboxing users; then, query the corresponding initial benchmark threshold, which is the mean value corresponding to the user group multiplied by the preset business target adjustment coefficient (determined by technical personnel in this field based on historical sample analysis), and can be adjusted dynamically. At the same time, according to the user grouping and value analysis table, the The threshold correction coefficient is set dynamically by technical personnel in this field based on core indicators and user grouping. For example, the repurchase rate of digital collections of high-value core users is 55%, and the repurchase rate of digital collections of new users is 10%. The preset business target adjustment coefficient is 1.03. The threshold correction coefficient for high-value core users is set to 1.1, then 55%x1.03x1.1=62.3% is the corresponding benchmark threshold, the purpose of which is to retain high-value users. The threshold correction coefficient for new users is set to 1.2, the purpose of which is to repurchase digital collections of new users. The rate base is small, a higher improvement coefficient is set, and the initial benchmark threshold is revised to determine the corresponding benchmark threshold. The obtained virtual blind box opening conversion rate, digital collection repurchase rate, unpacking process abandonment rate, consecutive days of not entering the scene, virtual booth stay time or collaborative unpacking frequency are compared with the corresponding benchmark threshold. Those that do not meet the threshold (virtual blind box opening conversion rate, digital collection repurchase rate, virtual booth stay time and collaborative unpacking frequency are less than the threshold, unpacking process abandonment rate and consecutive days of not entering the scene are greater than the threshold) are recorded and abnormal log record data is generated.
[0050] According to an embodiment of the present invention, the real-time indicator monitoring report and abnormal log record data are input into a preset behavior prediction model for processing to obtain individual user behavior prediction data and group user prediction trend data, including: Extracting data based on the user behavior profile data to obtain historical behavior characteristic data, including the average frequency of box opening and the average interval between repurchases of the collection; The box opening frequency mean value and the collection repeat purchase interval mean value, and the real-time index monitoring report and the abnormal log record data are input into a preset behavior prediction model for processing to obtain user individual behavior prediction data and group user prediction trend data.
[0051] It should be noted that, according to the box opening frequency mean value and the collection repeat purchase interval mean value of the reaction user historical behavior rule, the real-time index monitoring report and the abnormal log record data of the real-time behavior and the sudden factor are analyzed by the preset behavior prediction model to obtain user individual behavior prediction data (for example, the user A opening box probability is 80% in the future 7 days) focusing on the specific behavior of a certain user and group user prediction trend data (for example, in the future 7 days, the high-value core user group, the repeat purchase rate of the new online CIP series blind box is 60%) focusing on the user group, wherein the preset behavior prediction model is trained by inputting a large number of historical sample box opening frequency mean value and collection repeat purchase interval mean value, real-time index monitoring report and abnormal log record data and corresponding user individual behavior prediction data and group user prediction trend data into an LSTM neural network model under a TensorFlow framework.
[0052] According to the embodiment of the present application, the initial business scheme is optimized according to the user individual behavior prediction data and the group user prediction trend data to obtain an optimized business scheme, which comprises: According to the user individual behavior prediction data and the group user prediction trend data, the initial business scheme is optimized to obtain an optimized business scheme. The optimized business scheme comprises virtual scene optimization, operation strategy optimization and product function optimization.
[0053] It should be noted that the business scheme is realized according to the prediction result, virtual scene, operation strategy and product function optimization, the main purpose of virtual scene optimization is to improve the immersive interaction experience of users, and to reduce conversion anomalies, for example, if the group user prediction trend data shows that the longest stay time is at the BIP series exhibition stand, but the opening box conversion rate is low, and the user individual behavior prediction data indicates that 65% of high-value core users give up opening boxes due to the crowded interaction area of the exhibition stand, the virtual scene optimization scheme can be adjusted to adjust the visual layout of the BIP series virtual exhibition stand from 4 square centimeters to 9 square centimeters, and at the same time, improve the light effect to enhance the attraction to users; the main purpose of operation strategy optimization is to stimulate the consumption and interaction willingness of users according to user grouping, based on user individual behavior prediction data and group user prediction trend data, for example, user X has a 95% probability of opening a box in the next 3 days, and the preference is AIP hidden type, according to which the corresponding operation strategy is optimized, user D is pushed AIP hidden type blind box priority purchase right and appropriate discount, through group interaction amplification effect, the opening frequency and sharing conversion rate of users are improved; the purpose of product function optimization is to improve the practicability by solving abnormal points, for example, user individual behavior prediction data shows that 35% of users give up opening boxes due to the waiting process because the VR opening animation loading time exceeds 6 seconds, and the abnormal log record frequency is high, accordingly, product function optimization is performed, such as compressing the VR opening animation file size, and using preloading technology, the loading time is shortened from 6 seconds to 2 seconds, improving the virtual use experience and function stickiness of users.
[0054] The third aspect of the present application provides a readable storage medium, wherein a big data-based meta-universe user behavior analysis closed loop method program is stored in the readable storage medium, and the big data-based meta-universe user behavior analysis closed loop method program is executed by a processor to realize the steps of the big data-based meta-universe user behavior analysis closed loop method according to any one of the above.
[0055] The big data-based meta-universe user behavior analysis closed loop method, system and medium disclosed by the present application realize closed loop analysis of meta-universe user behavior through accurate data acquisition and efficient integration, behavior rule mining, dynamic monitoring and closed loop iterative optimization of adaptive scenes.
[0056] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other manners. The above described device embodiments are merely exemplary. For example, the division of the units is merely a logical function division. There can be another division manner for the actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed coupling, or direct coupling or communication connection between the components can be indirect coupling or communication connection through some interfaces, devices, or units, and can be electrical, mechanical, or in other forms.
[0057] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units; they can be located in one place, or distributed on multiple network units; and part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.
[0058] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in the form of hardware or in the form of hardware plus software functional units.
[0059] Those of ordinary skill in the art can understand that all or part of the steps of the above method embodiments can be completed by program instructions related to hardware, and the foregoing programs can be stored in a readable storage medium, and when the programs are executed, the steps of the above method embodiments are executed; and the foregoing storage medium includes mobile storage devices, read-only memories (ROMs), random access memories (RAMs), magnetic disks or optical disks, and various media that can store program codes.
[0060] Alternatively, the integrated units of the present application, if implemented in the form of software functional modules and sold or used as independent products, can also be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present application can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the embodiments of the present application. The foregoing storage medium includes mobile storage devices, ROMs, RAMs, magnetic disks or optical disks, and various media that can store program codes.
Claims
1. A closed-loop method for metaverse user behavior analysis based on big data, characterized by: The following steps are involved: Collect multimodal data of Metaverse users to obtain multimodal datasets, including virtual space operation record data, digital collection interaction record data, and user association data; Obtaining preset scenario-based data preprocessing standard rules, and preprocessing the virtual space operation record data, digital collection interaction record data, and user association data according to the preset scenario-based data preprocessing standard rules to obtain virtual space operation record standard data, digital collection interaction record standard data, and user association standard data; Performing data bit deviation correction processing on the virtual space operation record standard data, digital collection interaction record standard data, and user association standard data to generate user behavior archive data; Generate a business process funnel report and a user grouping and value analysis table based on the user behavior profile data through funnel model disassembly and user source tracing and grouping sorting; Analyze and process the user behavior profile data in combination with the business process funnel report and the user grouping and value analysis table to obtain real-time indicator monitoring reports and abnormal log record data; Input the real-time indicator monitoring report and abnormal log record data into a preset behavior prediction model for processing to obtain user individual behavior prediction data and group user prediction trend data; The initial business plan is optimized based on the individual user behavior prediction data and the group user prediction trend data to obtain an optimized business plan.
2. The closed-loop method for metaverse user behavior analysis based on big data according to claim 1 is characterized in that: The multimodal data of metaverse users is collected to obtain a multimodal dataset, including virtual space operation record data, digital collection interaction record data, and user association data, including: Collect multimodal data of Metaverse users to obtain multimodal datasets, including virtual space operation record data, digital collection interaction record data, and user association data; The virtual space operation record data includes movement trajectory, virtual booth coordinates, opening gesture dwell time and opening force; The digital collection interaction record data includes the digital blind box ID, the type of opening operation characteristic data, the frequency of opening the box, the amount of collection consumption, the collection viewing time record data, the number of shares and the virtual interaction rate; The user association data includes user virtual identity ID, user level and historical box opening IP preference.
3. The closed-loop method for metaverse user behavior analysis based on big data according to claim 2 is characterized in that: The step of obtaining preset scenario-based data preprocessing standard rules and preprocessing the virtual space operation record data, digital collection interaction record data, and user association data according to the preset scenario-based data preprocessing standard rules to obtain virtual space operation record standard data, digital collection interaction record standard data, and user association standard data includes: Obtain preset scenario-based data preprocessing standard rules, including data format unification rules, data cleaning rules, field definition unification rules, data association rules, and business adaptation rules; According to the preset scenario data preprocessing standard rules, the virtual space operation record data, digital collection interaction record data and user association data are preprocessed with timestamp unification, virtual coordinate conversion and abnormal data cleaning to obtain virtual space operation record standard data, digital collection interaction record standard data and user association standard data.
4. The closed-loop method for metaverse user behavior analysis based on big data according to claim 3 is characterized in that: The step of performing data bit deviation correction processing on the virtual space operation record standard data, the digital collection interaction record standard data, and the user association standard data to generate user behavior profile data includes: According to the user's virtual identity ID, the virtual space operation record standard data, the digital collection interaction record standard data and the user association standard data are associated with each other to generate initial user behavior profile data; The initial user behavior file data is subjected to a data bit deviation correction process for device timestamp deviation correction and virtual coordinate unification to generate user behavior file data.
5. The closed-loop method for metaverse user behavior analysis based on big data according to claim 4 is characterized in that: The user behavior profile data is disassembled through a funnel model and sorted by user tracing and grouping to generate a business process funnel report and a user grouping and value analysis table, including: According to the user behavior profile data, the conversion rate of each link is obtained by disassembling and analyzing the data through the funnel model; Comparing the conversion rate of the link with the preset standard conversion rate of the link by threshold; If the conversion rate of the link is lower than the preset standard conversion rate of the link, it is determined to be a user loss node; If the link conversion rate is greater than or equal to the preset link standard conversion rate, it is determined to be a normal node; Generate a business process funnel report based on the link conversion rate and the user churn node or normal node; Obtaining the user's registration duration based on the user's virtual identity ID; The user level, registration duration, box opening frequency, and historical box opening IP preference are input into a preset user grouping analysis model to obtain user groups, including new users, potential churn users, stable active users, high-value core users, or silent low-value users; Take the weighted sum of the collection consumption amount and virtual interaction rate corresponding to the user group to obtain the comprehensive value evaluation index corresponding to the user group, and sort them in descending order to generate a user group value ranking table; A user grouping and value analysis table is generated based on the user grouping and user group value ranking table.
6. The closed-loop method for metaverse user behavior analysis based on big data according to claim 5 is characterized in that: The analysis and processing based on the user behavior profile data combined with the business process funnel report and the user grouping and value analysis table to obtain real-time indicator monitoring reports and abnormal log record data includes: Data is extracted based on the user behavior profile data to obtain the virtual blind box opening conversion rate, digital collection repurchase rate, unboxing process abandonment rate, consecutive days of not entering the scene, virtual booth stay time and collaborative unboxing frequency corresponding to the user group; Generate a real-time indicator monitoring report based on the virtual blind box opening conversion rate, digital collection repurchase rate, unboxing process abandonment rate, consecutive days of not entering the scene, virtual booth stay time and collaborative unboxing frequency; Extract the initial benchmark thresholds corresponding to the virtual blind box opening conversion rate, digital collection repurchase rate, unboxing process abandonment rate, consecutive days of not entering the scene, virtual booth stay time, and collaborative unboxing frequency based on the business process funnel report; Determine the threshold correction coefficient based on the user segmentation and value analysis table, and correct the initial benchmark threshold to obtain the benchmark thresholds corresponding to the virtual blind box opening conversion rate, digital collection repurchase rate, unboxing process abandonment rate, consecutive days of not entering the scene, virtual booth stay time, and collaborative unboxing frequency; The virtual blind box opening conversion rate, digital collection repurchase rate, unboxing process abandonment rate, consecutive days of not entering the scene, virtual booth stay time or collaborative unboxing frequency are compared with the corresponding benchmark thresholds, and abnormal log record data is generated based on the threshold comparison results.
7. The closed-loop method for metaverse user behavior analysis based on big data according to claim 6 is characterized in that: The real-time indicator monitoring report and abnormal log record data are input into a preset behavior prediction model for processing to obtain individual user behavior prediction data and group user prediction trend data, including: Extracting data based on the user behavior profile data to obtain historical behavior characteristic data, including the average frequency of box opening and the average interval between repurchases of the collection; The average value of the box opening frequency and the average value of the collection repurchase interval, as well as the real-time indicator monitoring report and abnormal log record data are input into a preset behavior prediction model for processing to obtain user individual behavior prediction data and group user prediction trend data.
8. The closed-loop method for metaverse user behavior analysis based on big data according to claim 7 is characterized in that: Optimizing the initial business plan based on the individual user behavior prediction data and the group user prediction trend data to obtain an optimized business plan includes: Optimizing the initial business plan based on the individual user behavior prediction data and the group user prediction trend data to obtain an optimized business plan; The business optimization plan includes virtual scene optimization, operation strategy optimization and product function optimization.
9. A closed-loop system for analyzing user behavior in the metaverse based on big data, characterized by: The system comprises a memory and a processor, wherein the memory comprises a program of a closed-loop method for analyzing user behavior of a metaverse based on big data, and when the program of the closed-loop method for analyzing user behavior of a metaverse based on big data is executed by the processor, the following steps are implemented: Collect multimodal data of Metaverse users to obtain multimodal datasets, including virtual space operation record data, digital collection interaction record data, and user association data; Obtaining preset scenario-based data preprocessing standard rules, and preprocessing the virtual space operation record data, digital collection interaction record data, and user association data according to the preset scenario-based data preprocessing standard rules to obtain virtual space operation record standard data, digital collection interaction record standard data, and user association standard data; Performing data bit deviation correction processing on the virtual space operation record standard data, digital collection interaction record standard data, and user association standard data to generate user behavior archive data; Generate a business process funnel report and a user grouping and value analysis table based on the user behavior profile data through funnel model disassembly and user source tracing and grouping sorting; Analyze and process the user behavior profile data in combination with the business process funnel report and the user grouping and value analysis table to obtain real-time indicator monitoring reports and abnormal log record data; Input the real-time indicator monitoring report and abnormal log record data into a preset behavior prediction model for processing to obtain user individual behavior prediction data and group user prediction trend data; The initial business plan is optimized based on the individual user behavior prediction data and the group user prediction trend data to obtain an optimized business plan.
10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a closed-loop method program for metaverse user behavior analysis based on big data. When the closed-loop method program for metaverse user behavior analysis based on big data is executed by a processor, the steps of the closed-loop method for metaverse user behavior analysis based on big data as described in any one of claims 1 to 8 are implemented.
Citation Information
Cited By
User grouping analysis method and system oriented to differentiated operation
CN121458341A