A product recommendation method, apparatus, device, and medium

Abstract

A product recommendation method, device, equipment and medium are disclosed. The product recommendation method comprises: obtaining a regression error threshold; using an odd-even alternating regression prediction method to correct a current product recommendation sample set to obtain a current product recommendation corrected sample set according to the regression error threshold, the current product recommendation sample set and a current product recommendation model prediction result; training a product recommendation model based on the current product recommendation corrected sample set and returning to correct the current product recommendation sample set to obtain the current product recommendation corrected sample set until an iteration end condition is met, and outputting a model prediction result set; determining a target product recommendation model according to the model prediction result set, and performing product recommendation according to the target product recommendation model. The technical scheme of the embodiment of the present application can improve the prediction accuracy of the product recommendation model and improve the analysis ability of data rules and correlation.

Description

Technical Field

[0001] This invention relates to the field of model optimization technology, and in particular to a product recommendation method, apparatus, equipment, and medium. Background Technology

[0002] Currently, when using product recommendation models to predict product recommendations, the model's prediction performance is often poor and its accuracy cannot be guaranteed because the data attributes cannot well represent the true regression results. Furthermore, existing product recommendation models are not effective in analyzing data patterns and relationships. Summary of the Invention

[0003] This invention provides a product recommendation method, apparatus, device, and medium to address the problems of poor prediction performance, data pattern analysis, and correlation analysis in product recommendation models.

[0004] According to one aspect of the present invention, a product recommendation method is provided, comprising:

[0005] Obtain the regression error threshold;

[0006] Based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, the current product recommendation sample set is corrected by using the alternating odd and even regression prediction method to obtain the corrected sample set of the current product recommendation.

[0007] The product recommendation model is trained based on the current product recommendation correction sample set, and then the process is repeated until the iteration termination condition is met, and the model prediction result set is output. The model prediction result set is then output.

[0008] Based on the model prediction results set, determine the target product recommendation model, and make product recommendations based on the target product recommendation model.

[0009] According to another aspect of the present invention, a product recommendation device is provided, comprising:

[0010] The regression error threshold acquisition module is used to acquire the regression error threshold.

[0011] The corrected sample set acquisition module is used to correct the current product recommendation sample set based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, using an alternating odd-even regression prediction method, to obtain the current product recommendation corrected sample set.

[0012] The model prediction result set output module is used to train the product recommendation model based on the current product recommendation correction sample set, and return the operation of correcting the current product recommendation sample set by using an alternating odd and even regression prediction method according to the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, until the iteration termination condition is met, and output the model prediction result set.

[0013] The product recommendation module is used to determine the target product recommendation model based on the model prediction result set, and to make product recommendations based on the target product recommendation model.

[0014] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0015] At least one processor; and

[0016] A memory communicatively connected to the at least one processor; wherein,

[0017] The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the product recommendation method according to any embodiment of the present invention.

[0018] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the product recommendation method described in any embodiment of the present invention.

[0019] The technical solution of this invention involves obtaining a regression error threshold, and then, based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, using an alternating odd-even regression prediction method to correct the current product recommendation sample set, thus obtaining a corrected current product recommendation sample set. The product recommendation model is then trained based on this corrected sample set, and the process of correcting the current product recommendation sample set using an alternating odd-even regression prediction method, based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, is repeated until the iteration termination condition is met. Finally, a model prediction result set is output, and based on this set, a target product recommendation model is determined, and product recommendations are performed according to the target product recommendation model. In this scheme, by using an alternating odd-even regression prediction method and correcting the sample set with reference to the regression error threshold, the data attributes can better represent the true regression results, thereby reducing model validation bias and improving the accuracy of the product recommendation model. Furthermore, by continuously correcting the samples, the ability to discover data patterns and generalize can be expanded. This solves the problems of poor prediction performance, data pattern and correlation analysis performance of the product recommendation model, and improves the prediction accuracy of the product recommendation model and enhances the ability to analyze data patterns and correlations.

[0020] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart of a product recommendation method provided in Embodiment 1 of the present invention;

[0023] Figure 2 A flowchart of a product recommendation method provided in Embodiment 2 of the present invention;

[0024] Figure 3 This is a schematic diagram of the iterative optimization process of a product recommendation model provided in Embodiment 2 of the present invention;

[0025] Figure 4 This is a schematic diagram of the structure of a product recommendation device provided in Embodiment 3 of the present invention;

[0026] Figure 5A schematic diagram of an electronic device that can be used to implement embodiments of the present invention is shown. Detailed Implementation

[0027] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0028] It should be noted that the terms "target," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0029] Example 1

[0030] Figure 1 This is a flowchart of a product recommendation method provided in Embodiment 1 of the present invention. This embodiment is applicable to scenarios involving accurate prediction of product sales. The method can be executed by a product recommendation device, which can be implemented in hardware and / or software and can be configured in an electronic device. Figure 1 As shown, the method includes:

[0031] Step 110: Obtain the regression error threshold.

[0032] The regression error threshold can be a pre-set error threshold, and the specific value can be set by the user.

[0033] In this embodiment of the invention, a regression error threshold that meets the actual needs of regression analysis can be obtained.

[0034] Step 120: Based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, the current product recommendation sample set is corrected using the alternating odd-even regression prediction method to obtain the corrected sample set of the current product recommendation.

[0035] The current product recommendation sample set can be the set of samples currently input into the product recommendation model for model training and optimization. The current product recommendation model prediction result can be the result of the product recommendation model's predictive analysis of the currently input product recommendation sample set. Optionally, the current product recommendation model prediction result can include, but is not limited to, whether a user purchases and the quantity purchased. The product recommendation model can be any model with product recommendation functionality. The alternating odd-even regression prediction method can be a method where, after the product recommendation model completes its initial prediction, different regression prediction algorithms are used for odd-numbered and even-numbered iterations for optimization. The current product recommendation corrected sample set can be the sample set after correcting the current product recommendation sample set.

[0036] In this embodiment of the invention, the current product recommendation sample set can be input into the product recommendation model to obtain the prediction result of the current product recommendation model. Then, the prediction result of the current product recommendation model is compared with the label data in the current product recommendation sample set. Then, the samples in the current product recommendation sample set whose comparison error is greater than the regression error threshold are corrected according to the odd-even alternating regression prediction method to obtain the current product recommendation corrected sample set.

[0037] Step 130: Train the product recommendation model based on the current product recommendation correction sample set, and return to execute the operation of correcting the current product recommendation sample set by using the odd-even alternating regression prediction method according to the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, until the iteration termination condition is met, and output the model prediction result set.

[0038] The iteration termination condition can be a pre-set condition for ending the product recommendation model iteration. The model prediction result set can be a collection of prediction results output by the product recommendation model in each optimization iteration.

[0039] In this embodiment of the invention, the product recommendation model can be trained using the current product recommendation correction sample set to optimize the product recommendation model. The model then returns to the operation of correcting the current product recommendation sample set using an alternating odd-even regression prediction method based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model. When it is determined that the iterative operation of the product recommendation model meets the iteration termination condition, the model prediction result set is output.

[0040] Step 140: Based on the model prediction result set, determine the target product recommendation model, and make product recommendations based on the target product recommendation model.

[0041] The target product recommendation model can be a model determined based on the product recommendation model that matches the prediction results in the model prediction result set.

[0042] In this embodiment of the invention, each product recommendation model that matches the prediction results in the model prediction result set can be determined, that is, the product recommendation model optimized in each round of iteration. Then, the determined product recommendation models are combined to obtain the target product recommendation model, and product recommendation is performed using the target product recommendation model and the product data to be analyzed.

[0043] Optionally, the accuracy of the prediction results in the model prediction result set can be used as the weight of the corresponding product recommendation model, and the target product recommendation model can be determined in a balanced weight manner.

[0044] The technical solution of this invention involves obtaining a regression error threshold, and then, based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, using an alternating odd-even regression prediction method to correct the current product recommendation sample set, thus obtaining a corrected current product recommendation sample set. The product recommendation model is then trained based on this corrected sample set, and the process of correcting the current product recommendation sample set using an alternating odd-even regression prediction method, based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, is repeated until the iteration termination condition is met. Finally, a model prediction result set is output, and based on this set, a target product recommendation model is determined, and product recommendations are performed according to the target product recommendation model. In this scheme, by using an alternating odd-even regression prediction method and correcting the sample set with reference to the regression error threshold, the data attributes can better represent the true regression results, thereby reducing model validation bias and improving the accuracy of the product recommendation model. Furthermore, by continuously correcting the samples, the ability to discover data patterns and generalize can be expanded. This solves the problems of poor prediction performance, data pattern and correlation analysis performance of the product recommendation model, and improves the prediction accuracy of the product recommendation model and enhances the ability to analyze data patterns and correlations.

[0045] Example 2

[0046] Figure 2 This is a flowchart of a product recommendation method provided in Embodiment 2 of the present invention. This embodiment is a specific embodiment based on the above embodiment, and provides specific optional implementation methods for determining the target product recommendation model based on the model prediction result set. Figure 2 As shown, the method includes:

[0047] Step 210: Obtain the regression error threshold.

[0048] Step 220: Based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, the current product recommendation sample set is corrected using the alternating odd-even regression prediction method to obtain the corrected sample set of the current product recommendation.

[0049] In an optional embodiment of the present invention, after correcting the current product recommendation sample set using an alternating odd-even regression prediction method to obtain the corrected current product recommendation sample set, the method may further include: setting the hyperparameter range of the product recommendation model based on a greedy algorithm; and correcting the model hyperparameters of the product recommendation model based on the hyperparameter range, Optuna, and the current product recommendation sample set.

[0050] The hyperparameter range can be the range of parameters to be selected for the model's hyperparameters.

[0051] In this embodiment of the invention, a known greedy algorithm can be used to determine the range of hyperparameters of the product recommendation model. Then, based on Optuna (hyperparameter tuning framework), the model hyperparameters for modifying the product recommendation model can be determined from the range of hyperparameters according to the distribution of the current product recommendation sample set.

[0052] In an optional embodiment of the present invention, the current product recommendation sample set is corrected using an alternating odd-even regression prediction method based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, to obtain a corrected current product recommendation sample set. This may include: determining the error of each current product recommendation sample based on the current product recommendation sample set and the prediction results of the current product recommendation model; determining the product recommendation samples to be corrected and the product recommendation samples to be removed based on the regression error threshold and the errors of each current product recommendation sample; removing the product recommendation samples to be removed from the current product recommendation sample set; and correcting the product recommendation samples to be corrected using an alternating odd-even regression prediction method to obtain a corrected current product recommendation sample set.

[0053] The current product recommendation sample error characterizes the magnitude of the error between the prediction result of the current product recommendation model and the label data in the current product recommendation sample set. Product recommendation samples to be corrected are those whose error between the label data in the current product recommendation sample set and the prediction result of the current product recommendation model is greater than or equal to the regression error threshold. Product recommendation samples to be removed are those whose error between the label data in the current product recommendation sample set and the prediction result of the current product recommendation model is less than the regression error threshold.

[0054] In this embodiment of the invention, the error between the label data of each sample in the current product recommendation sample set and the prediction result of the current product recommendation model can be used as the error of each current product recommendation sample. Then, the regression error threshold is compared with the error of each current product recommendation sample. The samples corresponding to the current product recommendation sample errors that are greater than or equal to the regression error threshold are used as the product recommendation samples to be corrected, and the samples corresponding to the current product recommendation sample errors that are less than the regression error threshold are used as the product recommendation samples to be removed. The product recommendation samples to be removed in the current product recommendation sample set are further removed, and the product recommendation samples to be corrected are corrected by using an alternating odd and even regression prediction method to obtain the current product recommendation corrected sample set.

[0055] In an optional embodiment of the present invention, the method of alternating odd and even regression prediction to correct the product recommendation sample to be corrected may include: using a linear regression algorithm to correct the product recommendation sample to be corrected during odd-numbered rounds of model iteration; and using a nonlinear regression algorithm to correct the product recommendation sample to be corrected during even-numbered rounds of model iteration.

[0056] Odd-numbered rounds of model iteration can be performed on the product recommendation model after it has output its first prediction result. Even-numbered rounds of model iteration can be performed on the product recommendation model after it has output its first prediction result.

[0057] In this embodiment of the invention, after the product recommendation model outputs its initial prediction result, iterative optimization of the product recommendation model is initiated. During odd-numbered iterations, a linear regression algorithm is used to correct the product recommendation samples to be corrected, thereby using the corrected sample set (i.e., the current corrected product recommendation sample set) as the input to the product recommendation model during the next even-numbered iteration. During even-numbered iterations, a nonlinear regression algorithm is used to correct the product recommendation samples to be corrected, thereby using the corrected sample set as the input to the next odd-numbered iteration.

[0058] In an optional embodiment of the present invention, during odd-numbered rounds of model iteration, a linear regression algorithm is used to correct the product recommendation samples to be corrected. This may include: determining each local sample feature group of the product recommendation samples to be corrected based on the leave-one-out method; calculating the predicted features of each current sample based on the linear regression algorithm and each local sample feature group of the product recommendation samples to be corrected; and correcting the product recommendation samples to be corrected based on the predicted features of each current sample.

[0059] The local sample feature set can be a set of sample feature data obtained by removing features from the recommended product sample to be corrected based on the leave-one-out method. The current sample predicted features can be the sample features predicted by linear regression based on the local sample feature sets of the recommended product sample to be corrected.

[0060] In this embodiment of the invention, the leave-one-out method can be used to determine the local sample feature groups of each sample in at least one product recommendation sample to be corrected. Then, a linear regression algorithm is used to perform regression prediction on the corresponding local sample feature groups of at least one product recommendation sample to be corrected in sequence, so as to obtain the current sample prediction features that match each local sample feature group respectively, thereby correcting the corresponding product recommendation sample to be corrected according to the current sample prediction features.

[0061] Step 230: Train the product recommendation model based on the current product recommendation correction sample set, and return to execute the operation of correcting the current product recommendation sample set by using an alternating odd-even regression prediction method according to the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, until the iteration termination condition is met, and output the model prediction results.

[0062] In an optional embodiment of the present invention, the iteration termination condition may include: the current iteration number of the model is greater than the preset model iteration number, or the error of each current product recommendation sample is less than the regression error threshold.

[0063] The preset model iteration count can be the maximum number of model iterations set before the product recommendation model is iterated and optimized.

[0064] In this embodiment of the invention, before the product recommendation model is iterated, an iteration termination condition can be set. For example, if the current iteration number of the model is greater than the preset model iteration number, the training of the product recommendation model can be stopped. Alternatively, the training of the product recommendation model can be stopped when the error of each current product recommendation sample is less than the regression error threshold.

[0065] Step 240: Based on the model accuracy data of each iteration round matched with the model prediction result set, perform weight balancing on the product recommendation model of each iteration round to obtain the target product recommendation model, and make product recommendations based on the target product recommendation model.

[0066] Among them, the model accuracy data of the iteration round can describe the accuracy of the product recommendation model in the current iteration.

[0067] In this embodiment of the invention, the model accuracy data of the iteration round that matches the model prediction results in the model prediction result set can be determined. The model accuracy data of the iteration round is then used as the model weight of the product recommendation model in the corresponding iteration round. The product recommendation model is then multiplied with the matching model weight to complete the weight balancing process of the product recommendation model in each iteration round, thereby obtaining the target product recommendation model. The product data to be analyzed is then input into the target product recommendation model, and product recommendations are made based on the output results of the target product recommendation model.

[0068] Figure 3 This is a schematic diagram of the iterative optimization process of a product recommendation model provided in Embodiment 2 of the present invention, as shown below. Figure 3 As shown, it includes the following steps:

[0069] The initial product recommendation sample set input into the product recommendation model undergoes data cleaning, specifically outlier detection, data deduplication, and data interpolation, to avoid dirty data and skewed distributions that do not conform to model training. To ensure that data attributes effectively represent the target prediction results, further feature engineering operations can be performed. Specifically, based on actual business needs, enumerated features are quantified, some continuous features are discretized, and normalization is performed to eliminate dimensions and reduce errors, thereby controlling the magnitude of values ​​and improving the stability of the product recommendation model.

[0070] The product recommendation model is set as a lightweight and parallelizable LightGBM tree model, serving as the standard model for product recommendation model iteration. The hyperparameter search space of the LightGBM tree model is constructed. A greedy algorithm is used to set the hyperparameter range for the corresponding LightGBM tree model, and the Optuna framework is used to continuously adjust the hyperparameter values ​​based on the distribution of the sample set.

[0071] The system presets the number of model iterations and the regression error threshold, saves the product recommendation model prediction results output in each iteration during product recommendation model training, and identifies product recommendation samples with excessive errors that need to be corrected based on the regression error threshold.

[0072] The method of correcting the product recommendation samples to be corrected based on alternating odd-even regression prediction is as follows: After the model outputs its prediction result for the first time, it compares the prediction result with the product recommendation samples initially input into the model. Samples with errors greater than or equal to the regression error threshold are added to the product recommendation samples to be corrected. After the first iteration, the comparison operation between the model prediction result and the initial product recommendation samples is repeated to update the product recommendation samples to be corrected. In each iteration, the number of corrected samples gradually decreases, and the correction intensity gradually weakens to avoid data shift. Specifically, the method of alternating odd-even regression prediction is used to correct the products to be corrected. In odd-numbered model iterations, a linear regression algorithm is used. Specifically, when there are m features, one feature is used as the target variable, and the remaining m-1 features are used as input features to form a local training set. Linear regression is used to train m times to obtain m new columns of sample prediction features, which are added as new samples to the product recommendation samples to be corrected. In even-numbered model iterations, a nonlinear regression algorithm is used to obtain m new columns of sample prediction features. This process is repeated until the model iteration loop exits.

[0073] The multiple model predictions are then combined for a final comprehensive model prediction evaluation. Specifically, weights are assigned based on the error of each iteration's prediction, and a weighted average is taken as the final prediction.

[0074] The technical solution of this invention involves obtaining a regression error threshold, and then, based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, using an alternating odd-even regression prediction method to correct the current product recommendation sample set, thereby obtaining a corrected sample set. The product recommendation model is then trained based on this corrected sample set. The process continues until the iteration termination condition is met, at which point the model prediction results are output. Finally, based on the model accuracy data matched to the model prediction result set for each iteration, a weight balancing process is performed on the product recommendation model for each iteration to obtain a target product recommendation model. Finally, product recommendations are made based on the target product recommendation model. In this scheme, by using an alternating odd-even regression prediction method and correcting the sample set with reference to the regression error threshold, the data attributes can better represent the true regression results, thereby reducing model validation bias and improving the accuracy of the product recommendation model. Furthermore, by continuously correcting the samples, the ability to discover data patterns and generalize can be expanded. This solves the problems of poor prediction performance, data pattern and correlation analysis performance of the product recommendation model, and improves the prediction accuracy of the product recommendation model and enhances the ability to analyze data patterns and correlations.

[0075] Example 3

[0076] Figure 4 This is a schematic diagram of a product recommendation device provided in Embodiment 3 of the present invention. Figure 4 As shown, the device includes:

[0077] The regression error threshold acquisition module 310 is used to acquire the regression error threshold;

[0078] The corrected sample set acquisition module 320 is used to correct the current product recommendation sample set based on the regression error threshold, the current product recommendation sample set and the prediction results of the current product recommendation model, using an alternating odd-even regression prediction method, to obtain the current product recommendation corrected sample set.

[0079] The model prediction result set output module 330 is used to train the product recommendation model based on the current product recommendation correction sample set, and return the operation of correcting the current product recommendation sample set by using an alternating odd and even regression prediction method according to the regression error threshold, the current product recommendation sample set and the prediction result of the current product recommendation model, until the iteration termination condition is met, and output the model prediction result set.

[0080] The product recommendation module 340 is used to determine the target product recommendation model based on the model prediction result set, and to make product recommendations based on the target product recommendation model.

[0081] The technical solution of this invention involves obtaining a regression error threshold, and then, based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, using an alternating odd-even regression prediction method to correct the current product recommendation sample set, thus obtaining a corrected current product recommendation sample set. The product recommendation model is then trained based on this corrected sample set, and the process of correcting the current product recommendation sample set using an alternating odd-even regression prediction method, based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, is repeated until the iteration termination condition is met. Finally, a model prediction result set is output, and based on this set, a target product recommendation model is determined, and product recommendations are performed according to the target product recommendation model. In this scheme, by using an alternating odd-even regression prediction method and correcting the sample set with reference to the regression error threshold, the data attributes can better represent the true regression results, thereby reducing model validation bias and improving the accuracy of the product recommendation model. Furthermore, by continuously correcting the samples, the ability to discover data patterns and generalize can be expanded. This solves the problems of poor prediction performance, data pattern and correlation analysis performance of the product recommendation model, and improves the prediction accuracy of the product recommendation model and enhances the ability to analyze data patterns and correlations.

[0082] Optionally, the modified sample set acquisition module 320 includes a sample-to-be-processed determination unit and a modified sample set acquisition unit. The sample-to-be-processed determination unit is used to determine the error of each current product recommendation sample based on the current product recommendation sample set and the prediction results of the current product recommendation model, and to determine the product recommendation samples to be modified and the product recommendation samples to be removed based on the regression error threshold and the errors of each current product recommendation sample. The modified sample set acquisition unit is used to remove the product recommendation samples to be removed from the current product recommendation sample set, and to modify the product recommendation samples to be modified using an alternating odd-even regression prediction method to obtain the modified sample set for current product recommendations.

[0083] Optionally, the sample set acquisition unit includes an odd-round model acquisition subunit and an even-round model acquisition subunit. The odd-round model acquisition subunit is used to correct the product recommendation samples to be corrected using a linear regression algorithm during odd-round model iterations. The even-round model acquisition subunit is used to correct the product recommendation samples to be corrected using a non-linear regression algorithm during even-round model iterations.

[0084] Optionally, the odd-round model correction sample acquisition subunit is specifically used to determine each local sample feature group of the product recommendation sample to be corrected based on the leave-one-out method; calculate each current sample prediction feature according to the linear regression algorithm and each local sample feature group of the product recommendation sample to be corrected, and correct the product recommendation sample to be corrected based on each current sample prediction feature.

[0085] Optionally, the product recommendation device includes a model hyperparameter optimization module, used to set the hyperparameter range of the product recommendation model based on a greedy algorithm; and to correct the model hyperparameters of the product recommendation model based on the hyperparameter range, the hyperparameter tuning framework Optuna, and the current product recommendation sample set.

[0086] Optionally, the iteration termination condition includes: the current iteration number of the model is greater than the preset model iteration number, or the error of each current product recommendation sample is less than the regression error threshold.

[0087] Optionally, the product recommendation module 340 is specifically used to perform weight balancing processing on the product recommendation model of each iteration round based on the model accuracy data matched by the model prediction result set, so as to obtain the target product recommendation model.

[0088] The product recommendation device provided in this embodiment of the invention can execute the product recommendation method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of executing the method.

[0089] Example 4

[0090] Figure 5 A schematic diagram of an electronic device that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0091] like Figure 5 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0092] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0093] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as product recommendation methods.

[0094] In some embodiments, the product recommendation method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or mounted on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the product recommendation method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the product recommendation method by any other suitable means (e.g., by means of firmware).

[0095] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0096] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0097] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0098] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0099] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0100] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0101] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0102] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A product recommendation method, characterized in that, include: Obtain the regression error threshold; Based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model, the current product recommendation sample set is corrected by using an alternating odd-even regression prediction method to obtain the current product recommendation corrected sample set. The product recommendation model is trained based on the current product recommendation correction sample set, and then the operation of correcting the current product recommendation sample set by using an alternating odd-even regression prediction method according to the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model is returned, until the iteration termination condition is met, and the model prediction result set is output. Based on the model prediction result set, a target product recommendation model is determined, and product recommendations are made based on the target product recommendation model. The step of correcting the current product recommendation sample set using an alternating odd-even regression prediction method based on the regression error threshold, the current product recommendation sample set, and the prediction results of the current product recommendation model to obtain a corrected current product recommendation sample set includes: determining the error of each current product recommendation sample based on the current product recommendation sample set and the prediction results of the current product recommendation model; determining the product recommendation samples to be corrected and the product recommendation samples to be removed based on the regression error threshold and the errors of each current product recommendation sample; removing the product recommendation samples to be removed from the current product recommendation sample set; and correcting the product recommendation samples to be corrected using an alternating odd-even regression prediction method to obtain a corrected current product recommendation sample set. The method of using alternating odd-even regression prediction to correct the product recommendation sample includes: using a linear regression algorithm to correct the product recommendation sample during odd-numbered model iterations; and using a nonlinear regression algorithm to correct the product recommendation sample during even-numbered model iterations.

2. The method according to claim 1, characterized in that, During the odd-numbered rounds of model iteration, a linear regression algorithm is used to correct the product recommendation samples to be corrected, including: Based on the leave-one-out method, the feature groups of each local sample of the product recommendation sample to be corrected are determined; Based on the linear regression algorithm and the local sample feature groups of the product recommendation sample to be corrected, the predicted features of each current sample are calculated, and the product recommendation sample to be corrected is corrected based on the predicted features of each current sample.

3. The method according to claim 1, characterized in that, After correcting the current product recommendation sample set using the alternating odd-even regression prediction method to obtain the corrected current product recommendation sample set, the method further includes: Based on a greedy algorithm, the range of hyperparameters for the product recommendation model is set. Based on the hyperparameter range, the hyperparameter tuning framework Optuna, and the current product recommendation sample set, the model hyperparameters of the product recommendation model are corrected.

4. The method according to claim 1, characterized in that, The iteration termination conditions include: the current iteration number of the model is greater than the preset model iteration number, or the error of each current product recommendation sample is less than the regression error threshold.

5. The method according to claim 1, characterized in that, The step of determining the target product recommendation model based on the model prediction result set includes: Based on the model accuracy data of each iteration round matched with the model prediction result set, the product recommendation model of each iteration round is weighted and balanced to obtain the target product recommendation model.

6. A product recommendation device, characterized in that, include: The regression error threshold acquisition module is used to acquire the regression error threshold. The corrected sample set acquisition module is used to correct the current product recommendation sample set based on the regression error threshold, the current product recommendation sample set and the prediction result of the current product recommendation model, using an alternating odd-even regression prediction method, to obtain the current product recommendation corrected sample set. The model prediction result set output module is used to train the product recommendation model based on the current product recommendation correction sample set, and return to execute the operation of correcting the current product recommendation sample set by using an alternating odd-even regression prediction method according to the regression error threshold, the current product recommendation sample set and the prediction result of the current product recommendation model, until the iteration termination condition is met, and output the model prediction result set. The product recommendation module is used to determine a target product recommendation model based on the model prediction result set, and to recommend products based on the target product recommendation model. The modified sample set acquisition module includes a sample determination unit and a modified sample set acquisition unit; The sample determination unit is used to determine the error of each current product recommendation sample based on the current product recommendation sample set and the prediction result of the current product recommendation model, and to determine the product recommendation samples to be corrected and the product recommendation samples to be removed based on the regression error threshold and the error of each current product recommendation sample. The modified sample set acquisition unit is used to remove the product recommendation samples to be removed from the current product recommendation sample set, and to modify the product recommendation samples to be modified using an alternating odd-even regression prediction method to obtain the current product recommendation modified sample set. The sample set acquisition unit includes an odd-round model acquisition subunit and an even-round model acquisition subunit. The odd-round model acquisition subunit is used to correct the product recommendation samples to be corrected by using a linear regression algorithm during the odd-round model iteration. The even-round model correction sample acquisition sub-unit is used to correct the product recommendation samples to be corrected during even-round model iterations using a nonlinear regression algorithm.

7. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor to enable the at least one processor to perform the product recommendation method according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the product recommendation method according to any one of claims 1-5.

Images