E-commerce scene cold start recommendation method combining meta-learning and causal inference

Abstract

The invention relates to the technical field of e-commerce, in particular to an e-commerce scene cold start recommendation method combining meta-learning and causal inference, and the method comprises the steps: obtaining user and commodity feature data, and carrying out the vectorization representation; analyzing an interactive causal relationship between user features and commodity features, constructing a causal graph, and quantifying causal influence factors of user behaviors on commodity recommendation; constructing a cold start recommendation model based on a meta-learning framework, and converting a cold start problem into a cross-task transfer learning problem; training a meta-model, and learning a general mode between different tasks; using influence factors obtained by causal inference as constraint conditions, and optimizing a target function of the meta-learning recommendation model; and designing a personalized recommendation strategy, and generating a personalized cold start recommendation result in combination with causal inference constraints. According to the method, meta-learning and causal inference are comprehensively utilized, the cold start problem in an e-commerce scene is effectively solved, and recommendation individuation and interpretation are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of e-commerce, and specifically to an e-commerce scenario cold start recommendation method combining meta-learning and causal inference. Background Art

[0002] Existing e-commerce recommendation systems still have some limitations and shortcomings when facing the cold start problem. Traditional recommendation methods mainly rely on users' historical interaction data, and lack sufficient interaction information for new users or new products, resulting in low recommendation quality and personalization. Although some studies have tried to use auxiliary information of users and products to alleviate the cold start problem, they often fail to consider the intrinsic connection and influence mechanism between user characteristics and product characteristics, and ignore the causal relationship between them, making the recommendation results insufficiently interpretable and reasonable.

[0003] In addition, existing cross-domain recommendation methods aim to use the knowledge of the source domain to assist the recommendation task in the target domain, but they usually assume that user preferences and behavior patterns are consistent between different domains, ignoring the differences in user decision-making mechanisms in different scenarios, which limits the effect of transfer learning. At the same time, although existing recommendation methods can learn general knowledge from historical tasks and quickly adapt to new tasks, they mainly focus on the transfer ability of the model, and do not make full use of domain knowledge and causal relationships to guide the recommendation process, so that the robustness needs to be improved.

[0004] Therefore, there is an urgent need for a cold-start recommendation method that can comprehensively utilize multi-source data, consider causal mechanisms, adapt to different scenarios, and learn robust representations, so as to better solve the cold-start problem in e-commerce scenarios and provide more personalized and high-quality recommendation services.

[0005] In view of this, the present invention proposes an e-commerce scenario cold start recommendation method that combines meta-learning and causal inference. Summary of the invention

[0006] To achieve the above objectives, the present invention provides an e-commerce scenario cold start recommendation method combining meta-learning and causal inference. The specific technical scheme is as follows: The e-commerce scenario cold start recommendation method combining meta-learning and causal inference includes: Step 1: Obtain user feature data and product feature data, and quantize the extracted user and product features; Step 2: Analyze the interactive causal relationship between user feature data and product feature data, build a causal graph based on domain knowledge, and use causal inference methods to quantify the causal influence factors of user behavior on product recommendations; Step 3: Based on the meta-learning framework, a cold-start product recommendation model is constructed in the e-commerce sales scenario, and the cold-start problem is transformed into a cross-task transfer learning problem; Step 4: Train the metamodel based on historical task data so that the metamodel can learn common patterns between different tasks and adapt to cold start tasks in e-commerce sales scenarios; Step 5: Use the influencing factors obtained by causal inference as constraints to optimize the objective function of the meta-learning recommendation model; Step 6: Design personalized recommendation strategies based on the specific causal relationships of different users or products, and generate personalized cold start recommendation results based on the constraints of causal inference.

[0007] Preferably, obtaining user characteristic data includes collecting user demographic characteristics , behavioral characteristics and social features ; Obtain product feature data, including collecting product content features , Category Features and statistical characteristics .

[0008] Preferably, a multi-layer perceptron MLP model is constructed based on user characteristics. and product features As input, learn the embedding representation and ; The user embedding matrix is ​​obtained through the multi-layer perceptron MLP model and the product embedding matrix ,in, is the embedding dimension, For user collection, is a collection of products; each user and products Corresponding to one dimensional embedding vector and .

[0009] Preferably, a PC algorithm based on a graphical model is used to discover causal relationships and construct a causal graph; the steps of the PC algorithm are as follows: Step a: Initialize a completely undirected graph , where the node set Contains all user feature nodes and product feature nodes ; Step b: For any two nodes ,Through the conditional independence test, we can determine whether there is an edge between them; Step c: Assume that the condition set is ,like and In a given If independent under the condition of and The edge between Step d: Determine the threshold of the conditional independence test based on statistical knowledge ; Step e: Repeat steps b and c until the conditional independence test of all node pairs is completed; Step f: According to V-structure and domain knowledge, the undirected graph Orient some of the edges in to obtain a partial directed acyclic graph; Step g: Using the properties of acyclic graphs, direct the remaining undirected edges to obtain a completely directed acyclic graph, that is, a causal graph .

[0010] Preferably, the causal effect estimation of the frontier set is used, and for the causal graph Each node in ,Calculate the causal effect of each node on all child nodes; Set Node The set of all child nodes of , for each child node :Find from arrive All directed paths of ,in is the number of paths; For each path , calculate causal strength : ;in, Represents edge the causal strength of calculate right The causal effect : ; Combining the causal effects of all child nodes, we get the node The total causal effect : ; Introducing weight factors Weight the causal effect of user-product interaction to obtain the weighted causal effect :

[0011] in, , , They represent the causal effects of the three interactive modes of click, purchase and rating in e-commerce transaction scenarios, respectively. , and Indicates the factors affecting clicks, purchases, and ratings in e-commerce transaction scenarios; Quantify causal influence factors and calculate user characteristics based on weighted causal effects Product Features The causal influence factor :

[0012] All user characteristics For all product features The causal influence factor , integrate and generate the causal influence factor matrix , For all user feature nodes, It is the feature node of all commodities.

[0013] Preferably, the cold start recommendation problem in the e-commerce scenario is transformed into a cross-task transfer learning problem; Historical recommendation tasks , each task Contains a support set and a queryset , ; Support Set Contains user-item interaction data, represented as ,in Indicates feature vector of a user-item pair, Indicates whether the user is interested in the product. is the size of the support set; the query set Contains the user-item pair that needs to be predicted, expressed as ,in is the size of the query set.

[0014] Preferably, a model parameter meta-learning MAML framework is used to learn the initialization recommendation model parameters that are suitable for different tasks; Using historical task data Train the meta-model to learn common patterns between different tasks and adapt to new cold start tasks; the following meta-training process is used: Random initialization parameters , for each historical task , in the support set Calculate the gradient of the loss function; update the parameters using the gradient descent method , get task-specific parameters : In the query set Compute task-specific models The loss value of , repeat the meta-model training process until the initialization parameters Convergence or reaching the preset number of iterations.

[0015] Preferably, the impact factor matrix obtained by causal inference As a constraint, optimize the objective function of the meta-learning recommendation model; Design personalized recommendation strategies for specific causal relationships between different users or products; , according to the user Eigenvector and the causal influence factor matrix , calculate the user's preference weight vector for different product features :

[0016] in, is the user feature vector No. elements, , Indicates user and products Interaction score between .

[0017] Preferably, the constraints of causal inference are combined to generate personalized cold start recommendation results; For New Users , using user feature vector and product feature vector , through the adapted recommendation model , predict the preference scores of new users for all products: ; Based on the prediction score and the personalized preference weight vector , calculate the final product recommendation score for the cold start scenario :

[0018] in, is the balance factor, is the product feature vector No. elements; based on the cold start recommendation score Sort the products and generate new users A personalized cold-start product recommendation list.

[0019] An e-commerce scenario cold start recommendation system combining meta-learning and causal inference, which is used to implement the e-commerce scenario cold start recommendation method combining meta-learning and causal inference, including: a data collection module, a causal analysis module, a meta-learning module and a cold start recommendation module; The data collection module is used to obtain user feature data and product feature data, and quantize the extracted user and product features; The causal analysis module is used to analyze the interactive causal relationship between user feature data and product feature data, construct a causal graph based on domain knowledge, and use causal inference methods to quantify the causal influence factors of user behavior on product recommendations; The meta-learning module builds a cold-start product recommendation model in the e-commerce sales scenario based on the meta-learning framework, and transforms the cold-start problem into a cross-task transfer learning problem; trains the meta-model based on historical task data, so that the meta-model learns the common patterns between different tasks and adapts to the cold-start tasks in the e-commerce sales scenario; The cold start recommendation module is used to optimize the objective function of the meta-learning recommendation model by taking the influencing factors obtained by causal inference as constraints; design personalized recommendation strategies for the specific causal relationships of different users or commodities, and generate personalized cold start recommendation results in combination with the constraints of causal inference.

[0020] An electronic device comprises: a processor and a memory, wherein the memory stores a computer program that can be called by the processor; the processor executes the e-commerce scenario cold start recommendation method combining meta-learning and causal inference by calling the computer program stored in the memory.

[0021] A computer-readable storage medium stores instructions. When the instructions are executed on a computer, the computer executes the e-commerce scenario cold start recommendation method combining meta-learning and causal inference.

[0022] Beneficial effects of the present invention: The present invention obtains multi-dimensional features of users and products and vectorizes them, mines intrinsic semantic information, provides high-quality input for subsequent analysis and modeling, and improves recommendation accuracy and personalization.

[0023] The present invention introduces causal inference to analyze user-product interactions, quantifies the impact of user behavior on recommendations, reveals user decision-making logic, provides explainability for recommendations, and enhances persuasiveness and trust.

[0024] The present invention uses meta-learning to construct a cold start recommendation model, transforms cold start into cross-task transfer learning, learns general knowledge from historical tasks, quickly adapts to new scenarios, and alleviates the cold start problem.

[0025] The present invention trains a metamodel and learns commonalities between tasks, so that the metamodel can flexibly adapt to new cold start tasks, reduce data requirements and training costs, and improve practicality and generalization ability.

[0026] The present invention introduces causal inference influencing factors into the meta-learning optimization objectives, taking into account the causal constraints of user-product interactions, and generating recommendation results with stronger causal consistency and higher rationality.

[0027] The present invention designs a recommendation strategy based on the personalized causal relationship of users, generates personalized recommendations in combination with causal constraints, respects the causal preferences of users, and provides intimate and targeted recommendation services. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 A flow chart of the cold start recommendation method for e-commerce scenarios combining meta-learning and causal inference provided by the present invention; Figure 2 This is a structural diagram of the cold-start recommendation system for e-commerce scenarios that combines meta-learning and causal inference provided by the present invention. DETAILED DESCRIPTION

[0029] In order to better understand the present invention, a more detailed description will be made of various aspects of the present invention with reference to the accompanying drawings. It should be understood that these detailed descriptions are only descriptions of exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention in any way. Throughout the specification, the same reference numerals refer to the same elements. The expression "and / or" includes any and all combinations of one or more of the associated listed items.

[0030] In the accompanying drawings, the size, dimensions and shapes of the elements have been slightly adjusted for ease of illustration. The accompanying drawings are for illustration only and are not strictly drawn to scale. As used herein, the terms "substantially", "approximately" and similar terms are used as terms of approximation, not as terms of degree, and are intended to illustrate the inherent deviations in measurements or calculations that will be recognized by those of ordinary skill in the art. In addition, in the present invention, the order in which the steps are described does not necessarily represent the order in which these processes occur in actual operation, unless otherwise specified or can be derived from the context.

[0031] It should also be understood that expressions such as "comprises", "including", "having", "includes" and / or "comprising" are open rather than closed expressions in this specification, which indicate the presence of the stated features, elements and / or components, but do not exclude the presence of one or more other features, elements, components and / or combinations thereof. In addition, when expressions such as "at least one of..." appear after a list of listed features, they modify the entire list of features rather than just the individual elements in the list. In addition, when describing embodiments of the present invention, "may" is used to mean "one or more embodiments of the present invention". And, the term "exemplary" is intended to refer to an example or illustration.

[0032] Unless otherwise defined, all terms (including engineering terms and scientific and technological terms) used in this document have the same meaning as commonly understood by ordinary technicians in the field to which the present invention belongs. It should also be understood that unless otherwise clearly stated in the present invention, words defined in commonly used dictionaries should be interpreted as having the same meaning as their meaning in the context of the relevant technology, and should not be interpreted in an idealized or overly formal sense.

[0033] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0034] Example 1 Reference Figure 1 , which is the first embodiment of the present invention, provides an e-commerce scenario cold start recommendation method combining meta-learning and causal inference.

[0035] Step 1: Obtain user feature data and product feature data, and quantize the extracted user and product features.

[0036] Obtain user characteristic data, including collecting user demographic characteristics , behavioral characteristics and social features , where demographic characteristics Including gender, age and occupation, expressed as a Dimensional vector ; Behavioral characteristics Including browsing history, purchase history and rating data, represented as a Dimensional vector ; Social features Including friend relationships and interaction records, represented as a Dimensional vector .

[0037] Obtain product feature data, including collecting product content features , Category Features and statistical characteristics , where content features Includes title, description and image, represented as a Dimensional vector ; Category characteristics Including product categories and brands, represented as a Dimensional vector ; Statistical characteristics Including sales and ratings, expressed as a Dimensional vector .

[0038] Construct a multi-layer perceptron MLP model based on user characteristics and product features As input, learn its embedding representation and ;Optimize loss function ,in is the MLP network function, are network parameters, is the regularization coefficient.

[0039] The user embedding matrix is ​​obtained through the multi-layer perceptron MLP model and the product embedding matrix ,in is the embedding dimension; each user and products Corresponding to one dimensional embedding vector and The constructed embedding vector contains the compressed semantic information of the original features and is used for subsequent analysis and modeling.

[0040] Step 1 effectively captures the intrinsic characteristics of users and products by acquiring multi-dimensional feature data of users and products and learning their low-dimensional semantic representation, laying the foundation for subsequent causal analysis and recommendation modeling.

[0041] Step 2: Analyze the interactive causal relationship between user feature data and product feature data, build a causal graph based on domain knowledge, and use causal inference methods to quantify the causal impact factors of user behavior on product recommendations.

[0042] The PC algorithm based on the graph model is used to discover causal relationships and construct causal graphs; the basic steps of the PC algorithm are as follows: Step a: Initialize a completely undirected graph , where the node set Contains all user feature nodes and product feature nodes ; Step b: For any two nodes ,Through the conditional independence test, we can determine whether there is an edge between them; Step c: Assume that the condition set is ,like and In a given If independent under the condition of and The edge between Step d: Determine the threshold of the conditional independence test based on statistical knowledge ;Usually 0.05 or 0.01; Step e: Repeat steps b and c until the conditional independence test of all node pairs is completed; Step f: According to V-structure and domain knowledge, the undirected graph Orient some of the edges in to obtain a partial directed acyclic graph; Step g: Using the properties of acyclic graphs, direct the remaining undirected edges to obtain a completely directed acyclic graph, that is, a causal graph .

[0043] Using the causal effect estimation of the frontier set, for the causal graph Each node in ,Calculate the causal effect of each node on all child nodes.

[0044] Set Node The set of all child nodes of , for each child node :Find from arrive All directed paths of ,in is the number of paths.

[0045] For each path , calculate causal strength : ;in, Represents edge The causal strength of the hypothesis is estimated through a data-driven approach.

[0046] calculate right The causal effect : .

[0047] Combining the causal effects of all child nodes, we get the node The total causal effect : .

[0048] Introducing weight factors Weight the causal effect of user-product interaction to obtain the weighted causal effect :

[0049] in, , , They represent the causal effects of the three interactive modes of click, purchase and rating in e-commerce transaction scenarios, respectively. , and Represents the influencing factors of clicks, purchases, and ratings in e-commerce transaction scenarios.

[0050] Quantify causal influence factors and calculate user characteristics based on weighted causal effects Product Features The causal influence factor :

[0051] All user characteristics For all product features The causal influence factor , integrate and generate the causal influence factor matrix , used for optimization of subsequent recommendation models, For all user feature nodes, It is the feature node of all commodities.

[0052] Step 2 uses causal inference methods to analyze the interactive causal relationship between user characteristics and product characteristics, quantifies the influencing factors of user behavior on product recommendations, reveals the inherent causal mechanism of user-product interaction, and provides an explainable basis for personalized recommendations.

[0053] Step 3: Based on the meta-learning framework, a cold-start product recommendation model is constructed in the e-commerce sales scenario, and the cold-start problem is transformed into a cross-task transfer learning problem.

[0054] Convert the cold start recommendation problem in the e-commerce scenario into a cross-task transfer learning problem; obtain Historical recommendation tasks , each task Contains a support set and a queryset , .

[0055] Support Set Contains user-item interaction data, represented as ,in Indicates feature vector of a user-item pair, Indicates whether the user is interested in the product. is the size of the support set; the query set Contains the user-item pair that needs to be predicted, expressed as ,in is the size of the query set.

[0056] Step 3 builds a cold start recommendation model based on the meta-learning framework, transforming the cold start problem into a cross-task transfer learning problem, enabling the model to quickly adapt to new e-commerce scenarios, effectively alleviating the cold start problem and improving the generalization ability of the recommendation system.

[0057] Step 4: Train the meta-model based on historical task data so that the meta-model can learn the common patterns between different tasks and adapt to the cold start tasks in e-commerce sales scenarios.

[0058] The model parameter meta-learning MAML framework is used to learn the initialization recommendation model parameters that are suitable for different tasks; the goal of the parameter meta-learning MAML framework is to find the initialization parameters.

[0059] Using historical task data Train the meta-model to learn common patterns between different tasks and adapt to new cold start tasks; the following meta-training process is used: Random initialization parameters , for each historical task , in the support set Calculate the gradient of the loss function:

[0060] in, The parameters are The recommended model, is the cross entropy loss function, and the gradient descent method is used to update the parameters , get task-specific parameters : .

[0061] in, is the learning rate; in the query set Compute task-specific models The loss value:

[0062] Update initialization parameters ,in, is the meta-learning rate; repeat the meta-model training process until the initialization parameters Convergence or reaching the preset number of iterations.

[0063] Step 4 trains the meta-model on historical task data and learns common patterns between different tasks, so that the meta-model can quickly adapt to new cold-start tasks, greatly reducing the training cost in new scenarios and improving the flexibility and practicality of the recommendation system.

[0064] Step 5: Use the influencing factors obtained by causal inference as constraints to optimize the objective function of the meta-learning recommendation model.

[0065] The impact factor matrix obtained by causal inference As a constraint, optimize the objective function of the meta-learning recommendation model :

[0066] in, The parameters are The objective function of the recommendation model is is the balance factor, For the recommended model, connect User characteristics and The characteristic parameters of a product, express Norm.

[0067] Design personalized recommendation strategies for specific causal relationships between different users or products; , according to the user Eigenvector and the causal influence factor matrix , calculate the user's preference weight vector for different product features :

[0068] in, is the user feature vector No. elements, , Indicates user and products Interaction score between .

[0069] In step 5, the influencing factors obtained by causal inference are used as constraints to optimize the objective function of the meta-learning recommendation model, so that the model takes into account the causal relationship between user-product interaction while learning the general recommendation model, thereby improving the interpretability and rationality of the recommendation results.

[0070] Step 6: Design personalized recommendation strategies based on the specific causal relationships of different users or products, and generate personalized cold start recommendation results based on the constraints of causal inference.

[0071] For New Users , using user feature vector and product feature vector , through the adapted recommendation model , predict the preference scores of new users for all products: .

[0072] Based on the prediction score and the personalized preference weight vector , calculate the final product recommendation score for the cold start scenario :

[0073] in, is the balance factor, is the product feature vector No. elements; based on the cold start recommendation score Sort the products and generate new users A personalized cold-start product recommendation list.

[0074] Step 6 designs personalized recommendation strategies for different users, and generates a personalized cold start recommendation list based on the constraints of causal inference, which fully considers the user's personalized preferences and causal relationships, and improves the accuracy and satisfaction of recommendation results in cold start scenarios.

[0075] Example 2 Reference Figure 2 , which is the second embodiment of the present invention, provides an e-commerce scenario cold start recommendation system that combines meta-learning and causal inference.

[0076] The system includes: a data collection module, a causal analysis module, a meta-learning module and a cold start recommendation module.

[0077] The data collection module is used to obtain user feature data and product feature data, and quantize the extracted user and product features.

[0078] The causal analysis module is used to analyze the interactive causal relationship between user feature data and product feature data, construct a causal graph based on domain knowledge, and use a causal inference method to quantify the causal influence factor of user behavior on product recommendation.

[0079] The meta-learning module builds a cold-start product recommendation model in the e-commerce sales scenario based on the meta-learning framework, and transforms the cold-start problem into a cross-task transfer learning problem; the meta-model is trained based on historical task data so that the meta-model learns the common patterns between different tasks and adapts to the cold-start tasks in the e-commerce sales scenario.

[0080] The cold start recommendation module is used to optimize the objective function of the meta-learning recommendation model by taking the influencing factors obtained by causal inference as constraints; design personalized recommendation strategies for the specific causal relationships of different users or commodities, and generate personalized cold start recommendation results in combination with the constraints of causal inference.

[0081] Example 3 The present invention also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable codes, and when the computer-readable codes are executed by the one or more processors, the e-commerce scenario cold start recommendation method combining meta-learning and causal inference as described above may be executed.

[0082] The method or system according to the embodiment of the present invention may also be implemented with the aid of the architecture of the electronic device of the present invention.

[0083] An electronic device may include a bus, one or more CPUs, read-only memory (ROM), random access memory (RAM), a communication port for connecting to a network, input / output components, a hard disk, and the like.

[0084] A storage device in an electronic device, such as a ROM or a hard disk, can store the e-commerce scenario cold start recommendation method combining meta-learning and causal inference provided by the present invention.

[0085] A cold-start recommendation method for e-commerce scenarios that combines meta-learning with causal inference includes: obtaining user feature data and product feature data, and quantifying the extracted user and product features; analyzing the interactive causal relationship between user feature data and product feature data, building a causal graph based on domain knowledge, and using causal inference methods to quantify the causal impact factors of user behavior on product recommendations; building a cold-start product recommendation model in e-commerce sales scenarios based on a meta-learning framework, and converting the cold-start problem into a cross-task transfer learning problem; training a meta-model based on historical task data so that the meta-model learns common patterns between different tasks and adapts to cold-start tasks in e-commerce sales scenarios; using the impact factors obtained by causal inference as constraints to optimize the objective function of the meta-learning recommendation model; designing personalized recommendation strategies for specific causal relationships between different users or products, and generating personalized cold-start recommendation results based on the constraints of causal inference.

[0086] Furthermore, the electronic device may also include a user interface. Of course, the architecture of the present invention is only exemplary, and when implementing different devices, one or more components in the electronic device disclosed in the present invention may be omitted according to actual needs.

[0087] Example 4 The invention also discloses a computer-readable storage medium.

[0088] The computer-readable storage medium has computer-readable instructions stored thereon.

[0089] When the computer-readable instructions are executed by the processor, the e-commerce scenario cold start recommendation method combining meta-learning and causal inference according to the embodiments of the present invention described with reference to the above drawings can be executed.

[0090] The storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. In addition, according to an embodiment of the present invention, the process described above with reference to the flowchart may be implemented as a computer software program.

[0091] For example, the present invention provides a non-temporary machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided by the present invention, for example: obtaining user feature data and product feature data, and quantifying the extracted user and product features; analyzing the interactive causal relationship between user feature data and product feature data, constructing a causal graph in combination with domain knowledge, and using a causal inference method to quantify the causal impact factor of user behavior on product recommendation; constructing a cold-start product recommendation model in an e-commerce sales scenario based on a meta-learning framework, and converting the cold-start problem into a cross-task transfer learning problem; training a meta-model based on historical task data, so that the meta-model learns common patterns between different tasks and adapts to cold-start tasks in e-commerce sales scenarios; using the influencing factors obtained by causal inference as constraints to optimize the objective function of the meta-learning recommendation model; designing personalized recommendation strategies for specific causal relationships of different users or products, and generating personalized cold-start recommendation results in combination with the constraints of causal inference.

[0092] When the computer program is executed by the central processing unit (CPU), the above functions defined in the method of the present invention are performed. The method, apparatus, and device of the present invention may be implemented in many ways. For example, the method, apparatus, and device of the present invention may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware.

[0093] The above sequence for the steps of the method is for illustration only, and the steps of the method of the present invention are not limited to the sequence specifically described above unless otherwise specifically stated.

[0094] In addition, in some embodiments, the present invention can also be implemented as a program recorded in a recording medium, which includes machine-readable instructions for implementing the method according to the present invention. Therefore, the present invention also covers a recording medium storing a program for executing the method according to the present invention.

[0095] In addition, the parts of the above technical solutions provided in the embodiments of the present invention that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0096] The specific implementation modes as described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation mode of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A cold start recommendation method for e-commerce scenarios combining meta-learning and causal inference, characterized in that: include: Step 1: Obtain user feature data and product feature data, and quantize the extracted user and product features; Step 2: Analyze the interactive causal relationship between user feature data and product feature data, build a causal graph based on domain knowledge, and use causal inference methods to quantify the causal influence factors of user behavior on product recommendations; Step 3: Based on the meta-learning framework, a cold-start product recommendation model is constructed in the e-commerce sales scenario, and the cold-start problem is transformed into a cross-task transfer learning problem; Step 4: Train the metamodel based on historical task data so that the metamodel can learn common patterns between different tasks and adapt to cold start tasks in e-commerce sales scenarios; Step 5: Use the influencing factors obtained by causal inference as constraints to optimize the objective function of the meta-learning recommendation model; Step 6: Design personalized recommendation strategies based on the specific causal relationships of different users or products, and generate personalized cold start recommendation results based on the constraints of causal inference.

2. The e-commerce scenario cold start recommendation method combining meta-learning and causal inference according to claim 1 is characterized in that: Obtain user characteristic data, including collecting user demographic characteristics , behavioral characteristics and social features ; Obtain product feature data, including collecting product content features , Category Features and statistical characteristics .

3. The e-commerce scenario cold start recommendation method combining meta-learning and causal inference according to claim 2 is characterized in that: Construct a multi-layer perceptron MLP model based on user characteristics and product features As input, learn the embedding representation and ; The user embedding matrix is ​​obtained through the multi-layer perceptron MLP model and the product embedding matrix ,in, is the embedding dimension, For user collection, is a collection of products; each user and products Corresponding to one dimensional embedding vector and .

4. The e-commerce scenario cold start recommendation method combining meta-learning and causal inference according to claim 3 is characterized in that: The PC algorithm based on the graph model is used to discover causal relationships and construct causal graphs; the steps of the PC algorithm are as follows: Step a: Initialize a completely undirected graph , where the node set Contains all user feature nodes and product feature nodes ; Step b: For any two nodes ,Through the conditional independence test, we can determine whether there is an edge between them; Step c: Assume that the condition set is ,like and In a given If independent under the condition of and The edge between Step d: Determine the threshold of the conditional independence test based on statistical knowledge ; Step e: Repeat steps b and c until the conditional independence test of all node pairs is completed; Step f: According to V-structure and domain knowledge, the undirected graph Orient some of the edges in to obtain a partial directed acyclic graph; Step g: Using the properties of acyclic graphs, direct the remaining undirected edges to obtain a completely directed acyclic graph, that is, a causal graph .

5. The e-commerce scenario cold start recommendation method combining meta-learning and causal inference according to claim 4 is characterized in that: Using the causal effect estimation of the frontier set, for the causal graph Each node in ,Calculate the causal effect of each node on all child nodes; Set Node The set of all child nodes of , for each child node :Find from arrive All directed paths of ,in is the number of paths; For each path , calculate causal strength : ;in, Represents edge the causal strength of calculate right The causal effect : ; Combining the causal effects of all child nodes, we get the node The total causal effect : ; Introducing weight factors Weight the causal effect of user-product interaction to obtain the weighted causal effect : in, , , They represent the causal effects of the three interactive modes of click, purchase and rating in e-commerce transaction scenarios, , and Indicates the factors affecting clicks, purchases, and ratings in e-commerce transaction scenarios; Quantify causal influence factors and calculate user characteristics based on weighted causal effects Product Features The causal influence factor : All user characteristics For all product features The causal influence factor , integrate and generate the causal influence factor matrix , For all user feature nodes, It is the feature node of all commodities.

6. The e-commerce scenario cold start recommendation method combining meta-learning and causal inference according to claim 5 is characterized in that: Convert the cold start recommendation problem in the e-commerce scenario into a cross-task transfer learning problem; obtain Historical recommendation tasks , each task Contains a support set and a queryset , ; Support Set Contains user-item interaction data, represented as ,in Indicates feature vector of a user-item pair, Indicates whether the user is interested in the product. is the size of the support set; the query set Contains the user-item pair that needs to be predicted, expressed as ,in is the size of the query set.

7. The e-commerce scenario cold start recommendation method combining meta-learning and causal inference according to claim 6 is characterized in that: Adopt the model parameter meta-learning MAML framework to learn the initialization recommendation model parameters that are suitable for different tasks; Using historical task data Train the meta-model to learn common patterns between different tasks and adapt to new cold start tasks; the following meta-training process is used: Random initialization parameters , for each historical task , in the support set Calculate the gradient of the loss function on Update parameters using gradient descent , get task-specific parameters : In the query set Compute task-specific models The loss value of , repeat the meta-model training process until the initialization parameters Convergence or reaching the preset number of iterations.

8. The e-commerce scenario cold start recommendation method combining meta-learning and causal inference according to claim 7 is characterized in that: The impact factor matrix obtained by causal inference As a constraint, optimize the objective function of the meta-learning recommendation model; Design personalized recommendation strategies for specific causal relationships between different users or products; , according to the user Eigenvector and the causal influence factor matrix , calculate the user's preference weight vector for different product features : in, is the user feature vector No. elements, , Indicates user and products Interaction score between .

9. The e-commerce scenario cold start recommendation method combining meta-learning and causal inference according to claim 8, characterized in that: Combined with the constraints of causal inference, personalized cold start recommendation results are generated; For New Users , using user feature vector and product feature vector , through the adapted recommendation model , predict the preference scores of new users for all products: ; Based on the prediction score and the personalized preference weight vector , calculate the final product recommendation score for the cold start scenario : in, is the balance factor, is the product feature vector No. elements; based on the cold start recommendation score Sort the products and generate new users A personalized cold-start product recommendation list.

10. An e-commerce scenario cold start recommendation system combining meta-learning and causal inference, which is used to implement the e-commerce scenario cold start recommendation method combining meta-learning and causal inference as described in any one of claims 1 to 9, characterized in that: include: Data collection module, causal analysis module, meta-learning module and cold start recommendation module; The data collection module is used to obtain user feature data and product feature data, and quantize the extracted user and product features; The causal analysis module is used to analyze the interactive causal relationship between user feature data and product feature data, construct a causal graph based on domain knowledge, and use causal inference methods to quantify the causal influence factors of user behavior on product recommendations; The meta-learning module builds a cold-start product recommendation model in an e-commerce sales scenario based on a meta-learning framework, and transforms the cold-start problem into a cross-task transfer learning problem; Train the metamodel based on historical task data to enable the metamodel to learn common patterns between different tasks and adapt to cold start tasks in e-commerce sales scenarios; The cold start recommendation module is used to optimize the objective function of the meta-learning recommendation model by taking the influencing factors obtained by causal inference as constraints; design personalized recommendation strategies for the specific causal relationships of different users or commodities, and generate personalized cold start recommendation results in combination with the constraints of causal inference.

11. An electronic device, characterized in that: include: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; the processor executes the e-commerce scenario cold start recommendation method combining meta-learning and causal inference as described in any one of claims 1 to 9 by calling the computer program stored in the memory.

12. A computer-readable storage medium, characterized in that: Instructions are stored, and when the instructions are executed on a computer, the computer executes the e-commerce scenario cold start recommendation method combining meta-learning and causal inference as described in any one of claims 1 to 9.

Images