An interactive data management method and system based on artificial intelligence

Abstract

The present invention relates to the field of data management technology, and is an interactive data management method and system based on artificial intelligence. The specific method includes: real-time collection of historical order data and historical interaction data of users; intelligent analysis of historical order data and historical interaction data of users, determining the category of historical order products and predicting user satisfaction with historical order products; arranging the satisfaction in descending order and generating an initial product sequence; calculating and obtaining a real-time status correction factor; correcting the initial product sequence by the real-time status correction factor to obtain a first product sequence, synchronously obtaining real-time browsing interaction data of users, evaluating user browsing interest, and cyclically generating remaining product sequences. The present invention solves the problem in the prior art that the product recommendation system only relies on the user's historical purchase records or a single data source to generate recommendation results, resulting in low accuracy of recommended products and poor dynamics of the recommendation sequence.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and is an interactive data management method and system based on artificial intelligence. Background Art

[0002] At present, e-commerce platforms mainly rely on simple statistical methods or traditional recommendation systems when predicting user satisfaction, such as statistical analysis based on user ratings and purchase frequency. These methods can often only process limited data dimensions and fail to fully tap into users' comprehensive behavioral data, such as browsing history, click frequency, and shopping cart activities. Therefore, these traditional methods often fail to accurately reflect users' real needs and satisfaction, resulting in low prediction accuracy.

[0003] In addition, static models also fail to adapt to the dynamic changes in user behavior and personalized needs. In order to solve these problems, it is particularly important to develop a system based on artificial intelligence big models. Artificial intelligence big models can comprehensively consider multi-dimensional user behavior data and use deep learning and natural language processing technologies to provide more accurate and personalized satisfaction predictions. This system can adjust the prediction results in real time and dynamically adapt to changes in user needs, thereby significantly improving prediction accuracy and user experience.

[0004] In the existing disclosed invention technologies, for example, the patent with application publication number CN111553757A discloses a product recommendation method based on an adaptive neighborhood of a selection probability model. This method uses the user click frequency and the time of the last purchase (recency) in the user interaction behavior as initial data, establishes a selection probability model according to the binomial distribution to calculate the user's interest, optimizes the initial matrix and predicts the interest, and at the same time calculates the offset function based on the user deviation and the product deviation, and uses the stochastic gradient descent method to introduce a regularization term related to the difference between the user feature vector and its neighbor feature vector for adaptive updating.

[0005] The above patent cannot timely adjust the product recommendation sequence according to the real-time user interaction data and respond quickly to user behavior. It is not flexible enough in processing real-time user behavior and has the problems described in the background technology. Summary of the invention

[0006] The technical problem to be solved by the present invention is that in the prior art, the product recommendation system only relies on the user's historical purchase records or a single data source to generate recommendation results, resulting in low accuracy of recommended products and poor dynamics of the recommendation sequence. An interactive data management method and system based on artificial intelligence are proposed.

[0007] In order to achieve the above object, the technical solution of an interactive data management method based on artificial intelligence of the present invention comprises the following steps:

[0008] S1: Collect users’ historical order data and historical interaction data in real time through the front-end interface of the e-commerce platform;

[0009] S2: By setting up a user usage data prediction algorithm, the user's historical order data and historical interaction data are intelligently analyzed to determine the category of the historical order products and predict the user's satisfaction with the historical order products;

[0010] S3: extracting the user's satisfaction with historical order items, arranging the satisfaction in descending order, and generating an initial item sequence;

[0011] S4: obtaining the user's real-time status data, and importing the real-time status data into the real-time status correction factor calculation strategy to calculate and obtain the real-time status correction factor;

[0012] S5: Correcting the initial product sequence by the real-time state correction factor to obtain a first product sequence, synchronously obtaining real-time browsing interaction data of the user when browsing the first product sequence, and evaluating the user's browsing interest according to the real-time browsing interaction data of the user;

[0013] S6: adjusting the ratio of commodities in the first commodity sequence according to the browsing interest of the user, and cyclically generating the remaining commodity sequences.

[0014] Specifically, in S1, the historical order data specifically includes: basic information of historical orders, basic information of historical order products and historical order amount information; the historical interaction data includes: user's order creation path data, order browsing time data and similar product data in the shopping cart; among them, the user's order creation path data includes: order creation after searching keywords, order creation after browsing and order creation in the shopping cart.

[0015] Specifically, S2 includes the following specific steps:

[0016] S21: extract product display pictures based on the basic information of historical order products to form a historical order product data set;

[0017] S22: preprocessing the collected product display images, including usually normalizing and scaling pixel values ​​to a range of [0,1];

[0018] S23: Establish a product category screening model, and input the image data I into the input layer of the product category screening model. Apply multiple convolution kernels in the convolution layer to perform convolution operations on the input image to extract local features of the product display image. The convolution operation formula is:

[0019]

[0020] Among them, I is the input image data of the input layer, K is the convolution kernel, and F i,j is the convolution value of the convolution result at the pixel position (i, j); m is the row index of the convolution kernel, and n is the column index of the convolution kernel;

[0021] S24: Apply an activation function to perform a nonlinear transformation on the convolution result, where the activation function is a ReLU function;

[0022] S25: Flatten the pooled feature map and further process it through a fully connected layer to output the final feature vector;

[0023] S26: The extracted feature vector is used as input to the classifier in the product category screening model for training. During the training process, the labeled seasonal product data is used as supervision information to optimize the model parameters.

[0024] Specifically, S2 also includes the following specific steps:

[0025] S27: extracting historical interaction data, and fusing the historical interaction data to construct a feature vector set of historical interaction data, including: embedded order creation path data, browsing time data, and shopping cart data;

[0026] S28: Establishing a user satisfaction prediction model, wherein the user satisfaction prediction model includes: an input layer, a feature extraction layer, a feature fusion layer and a satisfaction prediction layer; the input layer is used to input a feature vector set of historical interaction data; the feature extraction layer extracts data features of order creation path data through a convolutional layer, extracts data features of browsing time data through a feedforward neural network, and extracts data features of shopping cart data through a recurrent neural network;

[0027] S29: Perform weighted fusion on the extracted data features and simultaneously use the gradient descent optimization algorithm to train the model;

[0028] S210: Perform the final satisfaction prediction through the fully connected layer.

[0029] Specifically, S3 includes the following specific steps:

[0030] S31: extracting user satisfaction with historical order items;

[0031] S32: Arrange the satisfaction scores in descending order and generate an initial product sequence, wherein the initial product sequence includes historical order products and similar products of the same category but different brands as the historical order products.

[0032] Specifically, S4 includes the following specific steps:

[0033] S41: Acquire real-time status data of the user, wherein the real-time status data includes real-time geographical location information of the user and ambient temperature information corresponding to the geographical location of the user;

[0034] S42: Importing the real-time status data into a real-time status correction factor calculation strategy to calculate and obtain a real-time status correction factor, wherein the real-time status correction factor calculation strategy is specifically as follows:

[0035]

[0036] Wherein, F is the real-time state correction factor;

[0037] α1 and α2 are the position proportional coefficient and the temperature proportional coefficient respectively;

[0038] r and k are both subscripts, r represents the amount of data in the user's real-time status data; k represents the amount of data in the last seasonal product order data in the historical order data;

[0039] wd r ,jd r ,te r Respectively represent the latitude information, longitude information and ambient temperature information of the user's geographical location information in the user's real-time status data;

[0040] wd k ,jd k ,te k Respectively represent the latitude information, longitude information and ambient temperature information of the user's geographical location information in the last seasonal product order data;

[0041] L TE They are the geographical location switching threshold and the ambient temperature switching threshold respectively.

[0042] Specifically, S5 includes the following specific steps:

[0043] S51: Configure the user status switching threshold. When the real-time status correction factor is greater than or equal to the user status switching threshold, correct the initial product sequence, add the top five products in the seasonal product list of the season in this period to the initial product sequence, and output the first product sequence; among which, the product category ratio of historical order products, similar products of the same category but different brands of historical order products, and seasonal products is 4:4:2;

[0044] When the real-time status correction factor is less than the user status switching threshold, the initial product sequence is directly output as the first product sequence;

[0045] S52: synchronously obtaining real-time browsing interaction data of the user when browsing the first product sequence;

[0046] S53: Importing the user's real-time browsing interaction data into the user interest evaluation strategy to evaluate the user's browsing interest.

[0047] Specifically, in S53, the user interest evaluation strategy is as follows:

[0048]

[0049] Among them, Q is the user's browsing interest;

[0050] s1 is the user browsing time factor. Furthermore, the user browsing time factor is the ratio of the user's actual browsing time to the expected standard browsing time, wherein the user's actual browsing time is the actual time the user stays on the current product page; the expected standard browsing time is the preset standard time the user stays on the product page;

[0051] s2 is the user browsing depth factor. Further, the user browsing depth factor is the position of the product currently browsed by the user when browsing the first product sequence and the total number of products in the first product sequence;

[0052] s3 is a user click frequency factor. Further, the user click frequency factor is a ratio of the number of clicks on a product by a user during browsing to the expected standard number of clicks, wherein the expected standard number of clicks is a preset standard number of clicks by a user on a product page;

[0053] s4 is the user scroll speed factor, Wherein, V is the average scrolling speed of the user when browsing the page, and v0 is the preset standard scrolling speed of the user when browsing the page;

[0054] a, b, c, d are all weight coefficients, a+b+c+d=1.

[0055] Specifically, the loop in S6 generates the remaining product sequence as follows:

[0056] S61: Standardize the calculated interest level to ensure that the interest level values ​​of all products are within a uniform range, where the standardized interest level Q 0 The calculation strategy is: Q u represents the user's browsing interest in the u-th item,

[0057] Q max ,Q min are the maximum and minimum browsing interest of the user for all products in the first product sequence, respectively;

[0058] S62: Calculate a target proportion of each product in the first product sequence according to the browsing interest of the user, where the product proportion is a coefficient of the standardized interest and the product category proportion;

[0059] S63: Adjust the product sequence according to the calculated product ratio, and put the products with high interest at the front.

[0060] In addition, the present invention provides an interactive data management system based on artificial intelligence, which includes the following modules:

[0061] Historical data collection module, satisfaction prediction module, initial product sequence generation module, correction factor calculation module, browsing interest calculation module and remaining product sequence generation module;

[0062] The historical data collection module collects the user's historical order data and historical interaction data in real time through the front-end interface of the e-commerce platform;

[0063] The satisfaction prediction module performs intelligent analysis on the user's historical order data and historical interaction data by setting a user usage data prediction algorithm, determines the category of the historical order goods and predicts the user's satisfaction with the historical order goods;

[0064] The initial product sequence generation module is used to extract the user's satisfaction with the historical order products, arrange the satisfaction in descending order, and generate an initial product sequence;

[0065] The correction factor calculation module is used to obtain the real-time status data of the user, and import the real-time status data into the real-time status correction factor calculation strategy to calculate and obtain the real-time status correction factor;

[0066] The browsing interest calculation module corrects the initial product sequence by the real-time state correction factor to obtain the first product sequence, synchronously obtains the real-time browsing interaction data of the user when browsing the first product sequence, and evaluates the user's browsing interest according to the real-time browsing interaction data of the user;

[0067] The remaining product sequence generation module adjusts the product ratio in the first product sequence according to the user's browsing interest, and cyclically generates the remaining product sequence.

[0068] A storage medium stores instructions, and when a computer reads the instructions, the computer executes the interactive data management method based on artificial intelligence.

[0069] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned interactive data management method based on artificial intelligence when executing the computer program.

[0070] Compared with the prior art, the technical effects of the present invention are as follows:

[0071] 1. The present invention can accurately judge the user's satisfaction with the product and improve the relevance of the recommendation by real-time collection and intelligent analysis of user historical orders and interaction data.

[0072] 2. The present invention utilizes the real-time status correction factor, and the system can dynamically adjust the recommended product sequence according to the user's latest browsing behavior to improve the real-time and accuracy of the recommendation. In addition, the present invention automatically generates a product recommendation sequence to reduce manual intervention and improve the system's automation level and product display efficiency.

[0073] 3. The present invention generates personalized product recommendations based on the user's historical behavior, improves user experience and satisfaction, and enhances user retention and engagement. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use 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 ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0075] in:

[0076] Figure 1 A schematic diagram of a flow chart of an interactive data management method based on artificial intelligence of the present invention;

[0077] Figure 2 The figure is a schematic diagram of the structure of an interactive data management system based on artificial intelligence of the present invention. DETAILED DESCRIPTION

[0078] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0079] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0080] Embodiment 1:

[0081] like Figure 1 As shown, an interactive data management method based on artificial intelligence in an embodiment of the present invention is as follows: Figure 1 As shown, the specific steps are as follows:

[0082] S1: Collect users’ historical order data and historical interaction data in real time through the front-end interface of the e-commerce platform;

[0083] In S1, the historical order data specifically includes: basic information of historical orders, basic information of historical order products and historical order amount information; the historical interaction data includes: user's order creation path data, order browsing time data and similar product data in the shopping cart; among them, the user's order creation path data includes: order creation after searching keywords, order creation after browsing and order creation in the shopping cart.

[0084] S2: By setting up a user usage data prediction algorithm, the user's historical order data and historical interaction data are intelligently analyzed to determine the category of the historical order products and predict the user's satisfaction with the historical order products;

[0085] S2 includes the following specific steps:

[0086] S21: extract product display pictures based on the basic information of historical order products to form a historical order product data set;

[0087] S22: preprocessing the collected product display images, including usually normalizing and scaling pixel values ​​to a range of [0,1];

[0088] S23: Establish a product category screening model, and input the image data I into the input layer of the product category screening model. Apply multiple convolution kernels in the convolution layer to perform convolution operations on the input image to extract local features of the product display image. The convolution operation formula is:

[0089]

[0090] Among them, I is the input image data of the input layer, K is the convolution kernel, and F i,j is the convolution value of the convolution result at the pixel position (i, j); m is the row index of the convolution kernel, and n is the column index of the convolution kernel;

[0091] S24: Apply an activation function to perform a nonlinear transformation on the convolution result, where the activation function is a ReLU function;

[0092] S25: Flatten the pooled feature map and further process it through a fully connected layer to output the final feature vector;

[0093] S26: The extracted feature vector is used as input to the classifier in the product category screening model for training. During the training process, the labeled seasonal product data is used as supervision information to optimize the model parameters.

[0094] S2 also includes the following specific steps:

[0095] S27: extracting historical interaction data, and fusing the historical interaction data to construct a feature vector set of historical interaction data, including: embedded order creation path data, browsing time data, and shopping cart data;

[0096] S28: Establishing a user satisfaction prediction model, wherein the user satisfaction prediction model includes: an input layer, a feature extraction layer, a feature fusion layer and a satisfaction prediction layer; the input layer is used to input a feature vector set of historical interaction data; the feature extraction layer extracts data features of order creation path data through a convolutional layer, extracts data features of browsing time data through a feedforward neural network, and extracts data features of shopping cart data through a recurrent neural network;

[0097] S29: Perform weighted fusion on the extracted data features and simultaneously use the gradient descent optimization algorithm to train the model;

[0098] Exemplarily, in this embodiment, the gradient descent optimization algorithm is the Adam optimizer;

[0099] S210: Perform the final satisfaction prediction through the fully connected layer.

[0100] S3: extracting the user's satisfaction with historical order items, arranging the satisfaction in descending order, and generating an initial item sequence;

[0101] S3 includes the following specific steps:

[0102] S31: extracting user satisfaction with historical order items;

[0103] S32: Arrange the satisfaction scores in descending order and generate an initial product sequence, wherein the initial product sequence includes historical order products and similar products of the same category but different brands as the historical order products.

[0104] S4: obtaining the user's real-time status data, and importing the real-time status data into the real-time status correction factor calculation strategy to calculate and obtain the real-time status correction factor;

[0105] S4 includes the following specific steps:

[0106] S41: Acquire real-time status data of the user, wherein the real-time status data includes real-time geographical location information of the user and ambient temperature information corresponding to the geographical location of the user;

[0107] S42: Importing the real-time status data into a real-time status correction factor calculation strategy to calculate and obtain a real-time status correction factor, wherein the real-time status correction factor calculation strategy is specifically as follows:

[0108]

[0109] Wherein, F is the real-time state correction factor;

[0110] α1 and α2 are the position proportional coefficient and the temperature proportional coefficient respectively;

[0111] r and k are both subscripts, r represents the amount of data in the user's real-time status data; k represents the amount of data in the last seasonal product order data in the historical order data;

[0112] wd r ,jd r ,te r Respectively represent the latitude information, longitude information and ambient temperature information of the user's geographical location information in the user's real-time status data;

[0113] wd k ,jd k ,te k Respectively represent the latitude information, longitude information and ambient temperature information of the user's geographical location information in the last seasonal product order data;

[0114] L TE They are the geographical location switching threshold and the ambient temperature switching threshold respectively.

[0115] S5: Correcting the initial product sequence by the real-time state correction factor to obtain a first product sequence, synchronously obtaining real-time browsing interaction data of the user when browsing the first product sequence, and evaluating the user's browsing interest according to the real-time browsing interaction data of the user;

[0116] S5 includes the following specific steps:

[0117] S51: Configure the user status switching threshold. When the real-time status correction factor is greater than or equal to the user status switching threshold, correct the initial product sequence, add the top five products in the seasonal product list of the season in this period to the initial product sequence, and output the first product sequence; among which, the product category ratio of historical order products, similar products of the same category but different brands of historical order products, and seasonal products is 4:4:2;

[0118] When the real-time status correction factor is less than the user status switching threshold, the initial product sequence is directly output as the first product sequence;

[0119] S52: synchronously obtaining real-time browsing interaction data of the user when browsing the first product sequence;

[0120] S53: Importing the user's real-time browsing interaction data into the user interest evaluation strategy to evaluate the user's browsing interest.

[0121] In S53, the user interest evaluation strategy is specifically as follows:

[0122]

[0123] Among them, Q is the user's browsing interest;

[0124] s1 is the user browsing time factor. Furthermore, the user browsing time factor is the ratio of the user's actual browsing time to the expected standard browsing time, wherein the user's actual browsing time is the actual time the user stays on the current product page; the expected standard browsing time is the preset standard time the user stays on the product page;

[0125] s2 is the user browsing depth factor. Further, the user browsing depth factor is the position of the product currently browsed by the user when browsing the first product sequence and the total number of products in the first product sequence;

[0126] s3 is a user click frequency factor. Further, the user click frequency factor is a ratio of the number of clicks on a product by a user during browsing to the expected standard number of clicks, wherein the expected standard number of clicks is a preset standard number of clicks by a user on a product page;

[0127] s4 is the user scroll speed factor, Wherein, V is the average scrolling speed of the user when browsing the page, and v0 is the preset standard scrolling speed of the user when browsing the page;

[0128] a, b, c, d are all weight coefficients, a+b+c+d=1.

[0129] Exemplarily, in this embodiment, the standard dwell time, the standard number of clicks, and the standard scrolling speed are all preset by those skilled in the art based on historical interaction data and actual technical requirements.

[0130] S6: adjusting the ratio of commodities in the first commodity sequence according to the browsing interest of the user, and cyclically generating the remaining commodity sequences.

[0131] The loop in S6 to generate the remaining product sequence is as follows:

[0132] S61: Standardize the calculated interest level to ensure that the interest level values ​​of all products are within a uniform range, where the standardized interest level Q 0 The calculation strategy is: Q u represents the user's browsing interest in the u-th item,

[0133] Q max ,Q minare the maximum and minimum browsing interest of the user for all products in the first product sequence, respectively;

[0134] S62: Calculate a target proportion of each product in the first product sequence according to the browsing interest of the user, where the product proportion is a coefficient of the standardized interest and the product category proportion;

[0135] S63: Adjust the product sequence according to the calculated product ratio, and put the products with high interest at the front.

[0136] Embodiment 2:

[0137] like Figure 2 As shown, an interactive data management system based on artificial intelligence in an embodiment of the present invention is as follows: Figure 2 As shown, it includes the following modules:

[0138] Historical data collection module, satisfaction prediction module, initial product sequence generation module, correction factor calculation module, browsing interest calculation module and remaining product sequence generation module;

[0139] The historical data collection module collects the user's historical order data and historical interaction data in real time through the front-end interface of the e-commerce platform;

[0140] The satisfaction prediction module performs intelligent analysis on the user's historical order data and historical interaction data by setting a user usage data prediction algorithm, determines the category of the historical order goods and predicts the user's satisfaction with the historical order goods;

[0141] The initial product sequence generation module is used to extract the user's satisfaction with the historical order products, arrange the satisfaction in descending order, and generate an initial product sequence;

[0142] The correction factor calculation module is used to obtain the real-time status data of the user, and import the real-time status data into the real-time status correction factor calculation strategy to calculate and obtain the real-time status correction factor;

[0143] The browsing interest calculation module corrects the initial product sequence by the real-time state correction factor to obtain the first product sequence, synchronously obtains the real-time browsing interaction data of the user when browsing the first product sequence, and evaluates the user's browsing interest according to the real-time browsing interaction data of the user;

[0144] The remaining product sequence generation module adjusts the product ratio in the first product sequence according to the user's browsing interest, and cyclically generates the remaining product sequence.

[0145] Embodiment three:

[0146] This embodiment provides an electronic device, including: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0147] The processor executes the above-mentioned interactive data management method based on artificial intelligence by calling the computer program stored in the memory.

[0148] The electronic device may have relatively large differences due to different configurations or performances, and may include one or more processors (Central Processing Units, CPU) and one or more memories, wherein the memory stores at least one computer program, which is loaded and executed by the processor to implement an interactive data management method based on artificial intelligence provided by the above method embodiment. The electronic device may also include other components for implementing the functions of the device, for example, the electronic device may also have components such as a wired or wireless network interface and an input and output interface to input and output data. This embodiment will not be described in detail here.

[0149] Embodiment 4:

[0150] This embodiment provides a computer-readable storage medium having a rewritable computer program stored thereon;

[0151] When the computer program runs on a computer device, the computer device executes the above-mentioned interactive data management method based on artificial intelligence.

[0152] For example, the computer readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, etc.

[0153] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When a computer instruction or computer program is loaded or executed on a computer, a process or function according to an embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center through a wired network or / and a wireless network. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0154] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0155] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0156] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only one, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0157] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0158] In summary, compared with the prior art, the technical effects of the present invention are as follows:

[0159] Compared with the prior art, the technical effects of the present invention are as follows:

[0160] 1. The present invention can accurately judge the user's satisfaction with the product and improve the relevance of the recommendation by real-time collection and intelligent analysis of user historical orders and interaction data.

[0161] 2. The present invention utilizes the real-time status correction factor, and the system can dynamically adjust the recommended product sequence according to the user's latest browsing behavior to improve the real-time and accuracy of the recommendation. In addition, the present invention automatically generates a product recommendation sequence to reduce manual intervention and improve the system's automation level and product display efficiency.

[0162] 3. The present invention generates personalized product recommendations based on the user's historical behavior, improves user experience and satisfaction, and enhances user retention and engagement.

[0163] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An interactive data management method based on artificial intelligence, characterized in that: The method comprises the following specific steps: S1: Collect users’ historical order data and historical interaction data in real time through the front-end interface of the e-commerce platform; S2: By setting up a user usage data prediction algorithm, the user's historical order data and historical interaction data are intelligently analyzed to determine the category of the historical order products and predict the user's satisfaction with the historical order products; S3: extracting the user's satisfaction with historical order items, arranging the satisfaction in descending order, and generating an initial item sequence; S4: obtaining the user's real-time status data, and importing the real-time status data into the real-time status correction factor calculation strategy to calculate and obtain the real-time status correction factor; S41: Acquire real-time status data of the user, wherein the real-time status data includes real-time geographical location information of the user and ambient temperature information corresponding to the geographical location of the user; S42: Importing the real-time status data into a real-time status correction factor calculation strategy to calculate and obtain a real-time status correction factor, wherein the real-time status correction factor calculation strategy is specifically as follows: Wherein, F is the real-time state correction factor; α1 and α2 are the position proportional coefficient and the temperature proportional coefficient respectively; r and k are both subscripts, r represents the amount of data in the user's real-time status data; k represents the amount of data in the last seasonal product order data in the historical order data; wd r ,jd r ,te r Respectively represent the latitude information, longitude information and ambient temperature information of the user's geographical location information in the user's real-time status data; wd k ,jd k ,te k Respectively represent the latitude information, longitude information and ambient temperature information of the user's geographical location information in the last seasonal product order data; L TE They are the geographical location switching threshold and the ambient temperature switching threshold respectively; S5: Correcting the initial product sequence by the real-time state correction factor to obtain a first product sequence, synchronously obtaining real-time browsing interaction data of the user when browsing the first product sequence, and evaluating the user's browsing interest according to the real-time browsing interaction data of the user; S6: adjusting the proportion of commodities in the first commodity sequence according to the browsing interest of the user, and cyclically generating the remaining commodity sequences; S61: Standardize the calculated interest level to ensure that the interest level values ​​of all products are within a uniform range, where the standardized interest level Q 0 The calculation strategy is: Q u represents the user’s browsing interest in the u-th item, Q max ,Q min are the maximum and minimum browsing interest of the user for all products in the first product sequence, respectively; S62: Calculate a target proportion of each product in the first product sequence according to the browsing interest of the user, where the product proportion is a coefficient of the standardized interest and the product category proportion; S63: Adjust the product sequence according to the calculated product ratio, and put the products with high interest at the front.

2. The interactive data management method based on artificial intelligence according to claim 1, characterized in that: In S1, the historical order data specifically includes: basic information of historical orders, basic information of historical order products and historical order amount information; the historical interaction data includes: the user's order creation path data, order browsing time data and similar product data in the shopping cart; among them, the user's order creation path data includes: order creation after searching keywords, order creation after browsing and order creation in the shopping cart.

3. The interactive data management method based on artificial intelligence according to claim 2, characterized in that: S2 includes the following specific steps: S21: extract product display pictures based on the basic information of historical order products to form a historical order product data set; S22: preprocessing the collected product display images, including usually normalizing and scaling pixel values ​​to a range of [0,1]; S23: Establish a product category screening model, and input the image data I into the input layer of the product category screening model. Apply multiple convolution kernels in the convolution layer to perform convolution operations on the input image to extract local features of the product display image. The convolution operation formula is: Among them, I is the input image data of the input layer, K is the convolution kernel, and F i,j is the convolution value of the convolution result at the pixel position (i, j); m is the row index of the convolution kernel, and n is the column index of the convolution kernel; S24: Apply an activation function to perform a nonlinear transformation on the convolution result, where the activation function is a ReLU function; S25: Flatten the pooled feature map and further process it through a fully connected layer to output the final feature vector; S26: The extracted feature vector is used as input to the classifier in the product category screening model for training. During the training process, the labeled seasonal product data is used as supervision information to optimize the model parameters.

4. The interactive data management method based on artificial intelligence according to claim 3, characterized in that: S2 also includes the following specific steps: S27: extracting historical interaction data, and fusing the historical interaction data to construct a feature vector set of historical interaction data, including: embedded order creation path data, browsing time data, and shopping cart data; S28: Establishing a user satisfaction prediction model, wherein the user satisfaction prediction model includes: an input layer, a feature extraction layer, a feature fusion layer and a satisfaction prediction layer; the input layer is used to input a feature vector set of historical interaction data; the feature extraction layer extracts data features of order creation path data through a convolutional layer, extracts data features of browsing time data through a feedforward neural network, and extracts data features of shopping cart data through a recurrent neural network; S29: Perform weighted fusion on the extracted data features and simultaneously use the gradient descent optimization algorithm to train the model; S210: Perform the final satisfaction prediction through the fully connected layer.

5. The interactive data management method based on artificial intelligence according to claim 4, characterized in that: S3 includes the following specific steps: S31: extracting user satisfaction with historical order items; S32: Arrange the satisfaction scores in descending order and generate an initial product sequence, wherein the initial product sequence includes historical order products and similar products of the same category but different brands as the historical order products.

6. The interactive data management method based on artificial intelligence according to claim 5, characterized in that: S5 includes the following specific steps: S51: Configure the user status switching threshold. When the real-time status correction factor is greater than or equal to the user status switching threshold, correct the initial product sequence, add the top five products in the seasonal product list of the season in this period to the initial product sequence, and output the first product sequence; among which, the product category ratio of historical order products, similar products of the same category but different brands of historical order products, and seasonal products is 4:4:2; When the real-time status correction factor is less than the user status switching threshold, the initial product sequence is directly output as the first product sequence; S52: synchronously obtaining real-time browsing interaction data of the user when browsing the first product sequence; S53: Importing the user's real-time browsing interaction data into the user interest evaluation strategy to evaluate the user's browsing interest.

7. The interactive data management method based on artificial intelligence according to claim 6, characterized in that: In S53, the user interest evaluation strategy is specifically as follows: Among them, Q is the user's browsing interest; s1 is the user browsing time factor. Furthermore, the user browsing time factor is the ratio of the user's actual browsing time to the expected standard browsing time, wherein the user's actual browsing time is the actual time the user stays on the current product page; the expected standard browsing time is the preset standard time the user stays on the product page; s2 is the user browsing depth factor. Further, the user browsing depth factor is the position of the product currently browsed by the user when browsing the first product sequence and the total number of products in the first product sequence; s3 is a user click frequency factor. Further, the user click frequency factor is a ratio of the number of clicks on a product by a user during browsing to the expected standard number of clicks, wherein the expected standard number of clicks is a preset standard number of clicks by a user on a product page; s4 is the user scroll speed factor, Wherein, V is the average scrolling speed of the user when browsing the page, and v0 is the preset standard scrolling speed of the user when browsing the page; a, b, c, d are all weight coefficients, a+b+c+d=1.

8. An interactive data management system based on artificial intelligence, which is implemented based on an interactive data management method based on artificial intelligence as claimed in any one of claims 1 to 7, characterized in that: The system includes the following modules: Historical data collection module, satisfaction prediction module, initial product sequence generation module, correction factor calculation module, browsing interest calculation module and remaining product sequence generation module; The historical data collection module collects the user's historical order data and historical interaction data in real time through the front-end interface of the e-commerce platform; The satisfaction prediction module performs intelligent analysis on the user's historical order data and historical interaction data by setting a user usage data prediction algorithm, determines the category of the historical order goods and predicts the user's satisfaction with the historical order goods; The initial product sequence generation module is used to extract the user's satisfaction with the historical order products, arrange the satisfaction in descending order, and generate an initial product sequence; The correction factor calculation module is used to obtain the real-time status data of the user, and import the real-time status data into the real-time status correction factor calculation strategy to calculate and obtain the real-time status correction factor; The browsing interest calculation module corrects the initial product sequence by the real-time state correction factor to obtain the first product sequence, synchronously obtains the real-time browsing interaction data of the user when browsing the first product sequence, and evaluates the user's browsing interest according to the real-time browsing interaction data of the user; The remaining product sequence generation module adjusts the product ratio in the first product sequence according to the user's browsing interest, and cyclically generates the remaining product sequence.

Images