Multi-touchpoint marketing effect attribution method based on AI and big data integration

Abstract

The present invention discloses a multi-touch marketing effect attribution method integrated with AI big data, which relates to the fields of information technology and data science. It is used to solve the problems of high computational redundancy, high resource consumption, and low execution efficiency of the overall model. By collecting user behavior data and transaction data of multiple touch points and performing preprocessing, a feature vector is constructed and the touch point weight coefficient is calculated. If the weight exceeds a set threshold, a dynamic time warping algorithm is used to preliminarily determine whether there is a causal relationship between the touch points. After determining that there is a causal relationship, the interaction frequency and depth between the touch points are collected, the degree of correlation is analyzed, and the correlation strength is calculated. According to the degree of correlation, a graph neural network algorithm is selected or the correlation strength is applied to a random forest algorithm to detect contact redundancy. This method improves the dynamic modeling capability of causal identification and realizes adaptive redundancy identification and optimization of the attribution model.

Description

Technical Field

[0001] The present invention relates to the fields of information technology and data science, and more specifically, to a multi-touchpoint marketing effect attribution method integrating AI and big data. Background Art

[0002] With the rapid development of big data and artificial intelligence technologies, multi-touch marketing attribution methods have become a crucial tool for companies to optimize marketing strategies and increase user conversion rates. However, existing multi-touch marketing attribution methods still have limitations in data integration, analytical accuracy, and real-time performance, making them unable to fully meet the needs of modern enterprises for efficient and precise marketing.

[0003] The existing technology has the following deficiencies:

[0004] At present, existing methods generally lack systematic feature screening strategies and graph structure optimization methods, and fail to effectively eliminate the noise interference brought by low-value touchpoints to the attribution model, which in turn leads to high computational redundancy of the overall model, high resource consumption, and low execution efficiency. Therefore, a multi-touchpoint marketing effect attribution method based on AI big data is proposed.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a multi-touch marketing effect attribution method based on AI big data fusion, which realizes accurate attribution of multi-touch marketing paths by integrating multi-source heterogeneous data and combining it with an intelligent analysis model to solve the problems raised in the above-mentioned background technology.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a multi-touch marketing effect attribution method based on AI big data fusion, comprising the following steps:

[0008] Step S1: Collect user behavior data and transaction data from multiple touchpoints and pre-process them;

[0009] Step S2: Construct a feature vector based on the preprocessed user behavior data and transaction data, and calculate the weight coefficient of each touchpoint. When the weight coefficient exceeds the set threshold, use the dynamic time warping algorithm to preliminarily determine whether there is a causal relationship between the touchpoints.

[0010] Step S3: When it is preliminarily determined that there is a causal relationship between the contacts, the interaction frequency and interaction depth between the contacts are collected, and the degree of association between the contacts is analyzed based on the interaction frequency and interaction depth. If the degree of association is high, the strength of the association between the contacts is calculated.

[0011] Step S4: Select a graph neural network algorithm based on the degree of correlation between the contacts or apply its correlation strength to the random forest algorithm to detect whether redundancy occurs between the contacts.

[0012] In a preferred embodiment, in step S1, the user behavior data of the touchpoints is represented by recording the number of clicks, dwell time, and page jump paths of users on different touchpoints and performing a hierarchical weighted summation to obtain the result;

[0013] The transaction data of the touchpoints is obtained by detecting the number of orders and transaction amounts completed by users at different touchpoints.

[0014] In a preferred embodiment, in step S1, the system sets up multiple groups of contact devices, detects user behavior data and transaction data of each group of contacts respectively, and performs normalization processing to obtain multiple groups of behavior data coefficients and transaction data coefficients.

[0015] In a preferred embodiment, in step S2, the processed behavior data coefficients are merged into a behavior data set, and the transaction data coefficients are merged into a transaction data set;

[0016] The specific steps for setting the weight threshold are as follows:

[0017] Perform the same processing on the two data sets, use the data set as the analysis data set, divide the analysis data set into two parts of equal size, calculate the average value of the two data sets respectively and compare them, and select the part with the larger average value as the new round of analysis data set;

[0018] Then, the new round of analysis data set is divided into two parts of equal size and the average value is calculated. The part with the smaller average value is selected as the new round of analysis data set;

[0019] Repeat the operation on the data set by alternating between taking the larger and smaller ones until the last data is left;

[0020] The average value of the data obtained from the processing of the two data sets is set as the weight threshold.

[0021] In a preferred embodiment, in step S2, the system calculates the behavior data coefficient and the transaction data coefficient of the touch point to be tested and takes the average value as the touch point weight coefficient;

[0022] When the contact weight coefficient exceeds the weight threshold, the dynamic time warping algorithm is used to perform preliminary matching on the time series data between contacts.

[0023] In a preferred embodiment, in step S3, the interaction frequency is obtained by obtaining the number of interactions between contacts through the timestamp field in the access log table; the interaction depth is reflected by recording the number of interaction levels between contacts in a set time period.

[0024] In a preferred embodiment, in step S3, fuzzy reasoning is used to analyze the degree of association between contacts. Specifically, the following steps are used: interaction frequency and interaction level are defined as input variables, which are divided into fuzzy sets;

[0025] The degree of association between contacts is defined as the output variable and divided into fuzzy sets; fuzzy rules are formulated to describe the impact of interaction frequency and interaction level on the degree of association between contacts;

[0026] Fuzzy reasoning is performed based on fuzzy rules to determine the degree of association between contacts.

[0027] In a preferred embodiment, in step S3, when calculating the association strength between contacts, the interaction frequency and the number of interaction levels between contacts are summed to obtain a normalized parameter, and the linear normalization formula is used to calculate and obtain the normalized parameter.

[0028] In a preferred embodiment, in step S4, when the degree of correlation between the contacts is low, the system uses a graph neural network algorithm to detect the relationship between the contacts;

[0029] When the degree of correlation between contacts is high, the system uses the random forest algorithm to detect the relationship between contacts;

[0030] Use the random forest algorithm to detect redundancy between contacts and give the redundancy strength between contacts;

[0031] When no redundancy problem occurs, the relationship between the contacts is recorded.

[0032] Technical effects and advantages of the present invention:

[0033] 1. The present invention collects user behavior data and transaction data from multiple contacts and preprocesses them, constructs a feature vector based on the preprocessed user behavior data and transaction data, calculates the weight coefficient of each contact, and compares it with the set threshold. When the weight coefficient exceeds the set threshold, the dynamic time warping algorithm is used to preliminarily determine whether there is a causal relationship between the contacts. When it is preliminarily determined that there is a causal relationship between the contacts, the interaction frequency and interaction depth between the contacts are collected, and the degree of correlation between the contacts is analyzed based on the interaction frequency and interaction depth; when the degree of correlation is high, the correlation strength between the contacts is calculated, and the graph neural network algorithm is selected according to the degree of correlation between the contacts, or its correlation strength is applied to the random forest algorithm to detect whether redundancy problems occur between the contacts. This improves the dynamic modeling capability of causal relationship identification between multiple contacts, and realizes the adaptive identification and optimization of contact redundancy problems by the attribution model. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1This is a schematic diagram of the overall process of the multi-touch marketing effect attribution method based on AI big data integration of the present invention.

[0035] Figure 2 This is a schematic diagram of the fuzzy reasoning process for analyzing the degree of correlation between touchpoints in the multi-touchpoint marketing effect attribution method based on AI big data fusion of the present invention. DETAILED DESCRIPTION

[0036] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0037] Example 1: The present invention provides a multi-touch marketing effect attribution method fused with AI big data. Figure 1 and attached Figure 2 Provide detailed explanation.

[0038] like Figure 1 As shown in the figure, the entire method process is completed by multiple modules working together, including data acquisition module, preprocessing module, feature vector construction module, weight calculation module, dynamic time warping module, interaction frequency analysis module, interaction depth analysis module, fuzzy reasoning module, association strength calculation module and algorithm selection module.

[0039] The following will describe the specific implementation steps of each module and the connection relationship and collaboration mode between them one by one in the order of the process.

[0040] First, the data collection module is responsible for acquiring user behavior and transaction data from multiple touchpoints, which can be websites, mobile apps, social media platforms, or other digital channels.

[0041] In actual applications, user behavior data of touchpoints cannot be obtained directly. Therefore, the system records the number of clicks, stay duration and page jump paths of users on different touchpoints and performs hierarchical weighted summation to obtain calculation results that can reflect user behavior characteristics.

[0042] Transaction data is obtained by detecting the number of orders and transaction amounts completed by users at different touch points.

[0043] To ensure the consistency and comparability of data, the system sets up multiple groups of contact devices to detect the user behavior data and transaction data of each group of contacts respectively, and transmits them to the preprocessing module for normalization.

[0044] The specific operation of normalization is to map the original data into a unified numerical range, thereby generating multiple sets of behavioral data coefficients and transaction data coefficients.

[0045] These coefficients serve as input data for subsequent modules and provide a basis for subsequent feature vector construction and weight calculation.

[0046] Next, the preprocessing module merges the processed multiple groups of behavior data coefficients and transaction data coefficients into a behavior data set and a transaction data set, respectively, and transmits them to the feature vector construction module.

[0047] The feature vector construction module constructs feature vectors based on the behavioral dataset and transaction dataset, and sets weight thresholds to evaluate the influence of each touchpoint.

[0048] The weight threshold setting process is as follows: for the two data sets, the same processing is performed on each data set, each data set is divided into two parts of equal size, the average values ​​of the two data sets are calculated and compared, and the part with the larger average value is selected as the new round of analysis data set;

[0049] Then, the new round of analysis data set is divided into two parts of equal size. After calculating the average value, the part with the smaller average value is selected as the new round of analysis data set.

[0050] Repeat this process by alternating between taking the larger and smaller values ​​until only one value remains. The average of the two processed values ​​is used as the final weight threshold.

[0051] Subsequently, the weight calculation module calculates the average value of the behavior data coefficient and the transaction data coefficient of the contact to be tested as the contact weight coefficient.

[0052] When the contact weight coefficient exceeds the weight threshold, the system determines that the contact has a high influence and transmits the relevant data to the dynamic time warping module for further analysis.

[0053] The dynamic time warping module uses the dynamic time warping algorithm to perform preliminary matching on the time series data between the contacts to be tested.

[0054] The dynamic time warping algorithm is an algorithm used to measure the similarity between two time series. It determines whether there is a causal relationship by calculating the optimal alignment path between the two time series.

[0055] In actual applications, the system takes the time series data between contacts as input, calculates the distance matrix between the two, and finds the optimal alignment path through dynamic programming methods.

[0056] If the cumulative distance of the optimal alignment path is less than the set threshold, the system determines that there is a causal relationship between the contacts and transmits the relevant data to the interaction frequency analysis module and the interaction depth analysis module for further analysis.

[0057] The interaction frequency analysis module obtains the number of interactions between contacts by accessing the timestamp field in the log table and records the interaction frequency between contacts.

[0058] The interaction depth analysis module records the number of interaction levels between touch points by setting a time period to reflect the depth of interaction.

[0059] The interaction frequency and the number of interaction levels are transmitted as input variables to the fuzzy reasoning module to analyze the degree of association between contacts.

[0060] like Figure 2 As shown, the fuzzy reasoning module uses fuzzy reasoning method to analyze the degree of association between contacts.

[0061] Specifically, the interaction frequency and the number of interaction levels are defined as input variables and divided into fuzzy sets;

[0062] The degree of association between contacts is defined as the output variable and divided into fuzzy sets;

[0063] Formulate fuzzy rules to describe the impact of interaction frequency and interaction level on the degree of association between touchpoints;

[0064] Fuzzy reasoning is performed based on fuzzy rules to determine the degree of association between contacts.

[0065] The result of fuzzy reasoning is transmitted to the association strength calculation module to calculate the association strength between contacts.

[0066] The association strength calculation module sums the interaction frequency and the number of interaction levels between contacts to obtain a normalized parameter, and uses a linear normalization formula based on the normalized parameter to calculate the association strength between contacts.

[0067] The calculation result of the association strength is transmitted to the algorithm selection module for selecting a suitable algorithm to detect the relationship between the contacts.

[0068] The algorithm selection module selects different algorithms for detection based on the degree of correlation between contacts. When the degree of correlation between contacts is low, the system uses a graph neural network algorithm to detect the relationship between contacts;

[0069] When the degree of correlation between contacts is high, the system uses the random forest algorithm to detect the relationship between contacts.

[0070] The random forest algorithm generates multiple decision tree models by sampling and training the data between contacts multiple times, and obtains the final detection results through a voting mechanism.

[0071] If the detection results show that there is a redundancy problem between the contacts, the system calculates the redundancy strength between the contacts; if there is no redundancy problem, the system records and stores the relationship between the contacts.

[0072] Throughout the entire process, each module is closely connected through data flow.

[0073] The data acquisition module transmits the collected raw data to the preprocessing module for normalization processing, and the preprocessing module transmits the processed data to the feature vector construction module and the weight calculation module.

[0074] The weight calculation module determines whether the contact has a high influence based on the calculation results and transmits the relevant data to the dynamic time warping module.

[0075] The matching results of the dynamic time warping module are transmitted to the interaction frequency analysis module and the interaction depth analysis module, and the analysis results of the interaction frequency analysis module and the interaction depth analysis module are transmitted to the fuzzy reasoning module.

[0076] The reasoning result of the fuzzy reasoning module is transmitted to the association strength calculation module, and the calculation result of the association strength calculation module is transmitted to the algorithm selection module.

[0077] The algorithm selection module selects an appropriate algorithm to detect the relationship between contacts according to the degree of association, and stores or feeds back the detection results to other modules.

[0078] The above embodiments illustrate the practical operation principle and process of the method of the present invention.

[0079] In actual application scenarios, for example, an e-commerce company wants to optimize its multi-touchpoint marketing strategy. This method can be used to collect and analyze user behavior data and transaction data from touchpoints such as its website, mobile application, and social media platforms.

[0080] First, the system uses the data collection module to obtain data such as the number of clicks, stay duration, page jump path, order quantity, and transaction amount of users on different touch points.

[0081] Then, the preprocessing module normalizes these data to generate behavioral data coefficients and transaction data coefficients.

[0082] The feature vector construction module constructs feature vectors based on these coefficients, and calculates the weight coefficient of each contact through the weight calculation module.

[0083] If the weight coefficient of a certain contact exceeds the set threshold, the dynamic time warping module matches the time series data between the contact and other contacts to determine whether there is a causal relationship.

[0084] If the match is successful, the system further analyzes the interaction frequency and interaction depth between the contacts, and calculates the degree of association and association strength between the contacts through the fuzzy reasoning module and the association strength calculation module.

[0085] Finally, the algorithm selection module selects an appropriate algorithm to detect the relationship between contacts based on the degree of association, and feeds back the detection results to the enterprise to help the enterprise optimize contact configuration and marketing strategy.

[0086] It can be seen from the above embodiments that the method of the present invention can effectively integrate multi-source heterogeneous data and combine it with an intelligent analysis model to achieve accurate attribution of multi-touchpoint marketing paths.

[0087] In practical applications, this method can not only reduce complex computing requirements, but also improve detection accuracy, clarify the relationship types between contacts, and provide enterprises with a scientific decision-making basis.

[0088] In order to better enable relevant personnel in this technical field to fully understand and implement the present invention, the specific implementation principle of the present invention is further supplemented below with reference to a specific application scenario.

[0089] In practical applications, suppose an e-commerce company wants to optimize its multi-touchpoint marketing strategy to improve user conversion rates and market competitiveness.

[0090] The enterprise collects and analyzes user behavior data and transaction data from its websites, mobile applications, social media platforms and other touchpoints through the method of the present invention.

[0091] The following are the specific operating steps and implementation principles of the method of the present invention in this scenario.

[0092] First, in the data collection module, the system deploys sensors or calls API interfaces to obtain data such as the number of user clicks, page dwell time, jump paths, order quantity, and transaction amount from multiple touchpoints.

[0093] This data reflects the user's behavioral characteristics and transaction preferences at different touchpoints. Since the original data may have inconsistent dimensions, the system transfers this data to the pre-processing module.

[0094] In the preprocessing module, the system uses a normalization algorithm to process the raw data and map it to a uniform numerical range.

[0095] For example, for the number of clicks and transaction amount, the system calculates their maximum and minimum values ​​and standardizes the data to coefficient values ​​between 0 and 1 using a linear transformation formula.

[0096] The normalized data is divided into behavioral dataset and transaction dataset and transmitted to the feature vector construction module respectively.

[0097] Subsequently, the feature vector construction module generates a feature vector based on the behavioral dataset and the transaction dataset, and sets a weight threshold to evaluate the influence of each touchpoint.

[0098] The process of setting the weight threshold is as follows: the system uses the behavioral dataset and the transaction dataset as analysis datasets respectively, and divides each dataset into two parts of equal size.

[0099] The system calculates the average of these two parts of data, compares their sizes, and selects the part with the larger average value as the new round of analysis data set.

[0100] Next, the system divides the new analysis data set into two parts again, calculates the average value, and selects the part with the smaller average value as the analysis data set for the next round.

[0101] This process repeats until only one value remains in each dataset. The system then calculates the average of the two final values ​​and uses it as the weight threshold.

[0102] The weight calculation module calculates the average value based on the behavioral data coefficient and transaction data coefficient of the touch point as the touch point weight coefficient.

[0103] If the weight coefficient of a contact exceeds the set threshold, the system determines that the contact has a high influence and transmits the relevant data to the dynamic time warping module.

[0104] The dynamic time warping module uses the dynamic time warping algorithm to match the time series data between the contacts to be tested.

[0105] Specifically, the system converts behavioral and transaction data between touchpoints into time series and calculates the distance matrix between them. Using dynamic programming, the system finds the optimal alignment path and records the cumulative distance.

[0106] If the cumulative distance of the optimal alignment path is less than the set threshold, the system determines that there is a causal relationship between the contacts.

[0107] For example, when a user browses a product on a social media platform and then visits an e-commerce website and completes a purchase within a short period of time, the system can identify this causal relationship and transmit the relevant data to the interaction frequency analysis module and the interaction depth analysis module.

[0108] The interaction frequency analysis module counts the number of interactions between contacts and records the interaction frequency by accessing the timestamp field in the log table.

[0109] For example, the system can count the number of times a user jumps from social media to an e-commerce platform. Meanwhile, the interaction depth analysis module records the number of interaction levels between touchpoints over a set time period, reflecting the depth of interaction.

[0110] For example, the system can record the number of intermediate touch points that the user passes through from the initial touch point to the final touch point. The interaction frequency and the number of interaction levels are transmitted as input variables to the fuzzy reasoning module.

[0111] like Figure 2 As shown in Figure 2, the fuzzy reasoning module uses fuzzy reasoning methods to analyze the degree of association between contacts. The system defines the interaction frequency and the number of interaction levels as input variables and divides them into fuzzy sets.

[0112] For example, the interaction frequency can be divided into three fuzzy sets: "low", "medium" and "high", and the interaction level can be divided into three fuzzy sets: "shallow", "medium" and "deep".

[0113] The degree of association between contacts is defined as an output variable and is also divided into fuzzy sets. The system formulates fuzzy rules to describe the impact of interaction frequency and interaction level on the degree of association between contacts.

[0114] For example, "When the interaction frequency is high and the number of interaction levels is deep, the degree of association between touch points is strong." After performing fuzzy reasoning based on fuzzy rules, the system determines the degree of association between touch points and transmits the result to the association strength calculation module.

[0115] The association strength calculation module obtains the normalized parameter based on the sum of the interaction frequency and the number of interaction levels between contacts, and uses the linear normalization formula to calculate the association strength between contacts.

[0116] For example, assuming that the interaction frequency between two touch points is 10 times and the number of interaction levels is 3, the system sums these two values ​​to obtain a normalization parameter 13, and maps it to an association strength value between 0 and 1 through the normalization formula.

[0117] This result is transmitted to the algorithm selection module.

[0118] The algorithm selection module selects the appropriate algorithm for detection based on the degree of correlation between the contacts. When the degree of correlation between the contacts is low, the system uses a graph neural network algorithm to detect the relationship between the contacts.

[0119] The graph neural network algorithm constructs a graph structure model between contacts and analyzes the connection patterns between nodes to identify whether there are potential connections between contacts.

[0120] When the degree of correlation between contacts is high, the system uses the random forest algorithm for detection.

[0121] The random forest algorithm generates multiple decision tree models by sampling and training the data between contacts multiple times, and obtains the final detection results through a voting mechanism.

[0122] If the test results show that there is redundancy between contacts, the system will further calculate the redundancy strength between contacts;

[0123] If no redundancy problem occurs, the system records and stores the relationship between the contacts.

[0124] It can be seen from the above steps that the method of the present invention can effectively integrate multi-source heterogeneous data and combine it with an intelligent analysis model to achieve accurate attribution of multi-touchpoint marketing paths.

[0125] In practical applications, this method can not only reduce complex computing requirements, but also improve detection accuracy, clarify the relationship types between contacts, and provide enterprises with a scientific decision-making basis.

[0126] For example, if analysis reveals that the correlation strength between certain touchpoints is low and there is redundancy, companies can adjust resource allocation and reduce investment in redundant touchpoints, thereby optimizing marketing strategies and improving market competitiveness.

[0127] Specifically, the above content describes the multi-touch marketing effect attribution method of AI big data fusion of the present invention from the perspective of a modular system. Furthermore, the multi-touch marketing effect attribution method of AI big data fusion includes the following steps:

[0128] Step S1: Collect user behavior data and transaction data from multiple touchpoints and pre-process them;

[0129] Step S2: Construct a feature vector based on the preprocessed user behavior data and transaction data, and calculate the weight coefficient of each touchpoint. When the weight coefficient exceeds the set threshold, use the dynamic time warping algorithm to preliminarily determine whether there is a causal relationship between the touchpoints.

[0130] Step S3: When it is preliminarily determined that there is a causal relationship between the contacts, the interaction frequency and interaction depth between the contacts are collected, and the degree of association between the contacts is analyzed based on the interaction frequency and interaction depth. If the degree of association is high, the strength of the association between the contacts is calculated.

[0131] Step S4: Select a graph neural network algorithm based on the degree of correlation between the contacts or apply its correlation strength to the random forest algorithm to detect whether redundancy occurs between the contacts.

[0132] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0133] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0134] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0135] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0136] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0137] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0138] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0139] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0140] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0141] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. The multi-touchpoint marketing effect attribution method based on AI and big data integration is characterized by: The following steps are involved: Step S1: Collect user behavior data and transaction data from multiple touchpoints and pre-process them; Step S2: Construct a feature vector based on the preprocessed user behavior data and transaction data, and calculate the weight coefficient of each touchpoint. When the weight coefficient exceeds the set threshold, use the dynamic time warping algorithm to preliminarily determine whether there is a causal relationship between the touchpoints. Step S3: When it is preliminarily determined that there is a causal relationship between the touchpoints, the interaction frequency and interaction depth between the touchpoints are collected, and the degree of correlation between the touchpoints is analyzed based on the interaction frequency and interaction depth; When the degree of association is high, the association strength between contacts is calculated; In step S3, the interaction frequency is obtained by accessing the timestamp field in the log table to obtain the number of interactions between contacts; the interaction depth is reflected by recording the number of interaction levels between contacts in a set time period; In step S3, fuzzy reasoning is used to analyze the degree of association between contacts. The specific steps are as follows: interaction frequency and interaction level are defined as input variables and divided into fuzzy sets; The degree of association between contacts is defined as the output variable and divided into fuzzy sets; fuzzy rules are formulated to describe the impact of interaction frequency and interaction level on the degree of association between contacts; Perform fuzzy reasoning based on fuzzy rules to determine the degree of association between contacts; Step S4: Select a graph neural network algorithm based on the degree of association between contacts or apply the association strength between contacts to a random forest algorithm to detect whether redundancy occurs between contacts.

2. The multi-touchpoint marketing effect attribution method based on AI big data fusion according to claim 1 is characterized by: In step S1, the user behavior data of the touchpoints is represented by recording the number of clicks, stay time, and page jump paths of users on different touchpoints and performing a hierarchical weighted summation to obtain the calculation result; The transaction data of the touchpoints is obtained by detecting the number of orders and transaction amounts completed by users at different touchpoints.

3. The multi-touch marketing effect attribution method based on AI big data fusion according to claim 2 is characterized by: In step S1, the system sets up multiple groups of contact devices, detects user behavior data and transaction data of each group of contacts respectively, and performs normalization processing to obtain multiple groups of behavior data coefficients and transaction data coefficients.

4. The multi-touch marketing effect attribution method based on AI big data fusion according to claim 3 is characterized by: In step S2, the processed behavior data coefficients are merged into a behavior data set, and the transaction data coefficients are merged into a transaction data set; The specific steps for setting the weight threshold are as follows: Perform the same processing on the two data sets, use the data set as the analysis data set, divide the analysis data set into two parts of equal size, calculate the average value of the two data sets respectively and compare them, and select the part with the larger average value as the new round of analysis data set; Then, the new round of analysis data set is divided into two parts of equal size and the average value is calculated. The part with the smaller average value is selected as the new round of analysis data set; Repeat the operation on the data set by alternating between taking the larger and smaller ones until the last data is left; The average value of the data obtained from the processing of the two data sets is set as the weight threshold.

5. The multi-touchpoint marketing effect attribution method based on AI big data fusion according to claim 4 is characterized by: In step S2, the system calculates the behavior data coefficient and transaction data coefficient of the touch point to be tested and takes the average value as the touch point weight coefficient; When the contact weight coefficient exceeds the weight threshold, the dynamic time warping algorithm is used to perform preliminary matching on the time series data between contacts.

6. The multi-touch marketing effect attribution method based on AI big data fusion according to claim 1 is characterized by: In step S3, when calculating the association strength between contacts, the interaction frequency and the number of interaction levels between contacts are summed to obtain a normalized parameter, and the linear normalization formula is used to calculate and obtain the normalized parameter.

7. The multi-touchpoint marketing effect attribution method based on AI big data fusion according to claim 6 is characterized by: In step S4, when the degree of correlation between contacts is low, the system uses a graph neural network algorithm to detect the relationship between contacts; When the degree of correlation between contacts is high, the system uses the random forest algorithm to detect the relationship between contacts; Use the random forest algorithm to detect redundancy between contacts and give the redundancy strength between contacts; When no redundancy problem occurs, the relationship between the contacts is recorded.

Images