Commodity pushing method and device based on AI and deep learning

Abstract

The invention relates to the technical field of commodity pushing, and discloses a commodity pushing method and device based on AI and deep learning. The method comprises the steps of configuring a first terminal of the retail scale, performing commodity pushing by using the first terminal and a preset public pushing model, obtaining a second terminal of a user, and performing commodity interaction by using the second terminal and a preset personal interaction model. The device corresponds to the method. According to the application, the display problem of the push information is solved through the preset public screen, the commodity searching problem is solved through processing the position of the commodity data, the data security problem is solved through the preset protocol information, and the intelligent improvement of the retail scale is realized by using the first terminal. The technical awkward situation of lack of linkage with the personal intelligent terminal of the user is improved, effective processing of commodity data is realized through the public pushing model and the personal interaction model, and development of intelligent commodity pushing is promoted. And finally, commodity pushing based on AI and deep learning is realized.

Description

Technical Field

[0001] The present application relates to the technical field of product push, and specifically to a product push method and device based on AI and deep learning. Background Art

[0002] With the popularization of AI technology in smart terminals and the gradual expansion of personal data, product selection based on AI technology has also gradually become popular. During the popularization process, the inventor discovered the following technical problems:

[0003] 1. When shopping in a mall, users do not hold their mobile phones and look at the screen to choose products. In the foreseeable future, even if they choose products through the screen, the process will be inconvenient;

[0004] 2. The difficulty of finding goods. When you want to find a certain item, it takes a long time to ask the counter staff for the more obscure items, or even the counter staff do not know where to find them;

[0005] 3. The issue of data security: In order to achieve personalized customization for users, it is necessary to obtain users’ private data, which poses obvious security issues when transmitted on the public network.

[0006] Retail scales are common hardware equipment in shopping malls. With the development of AI technology, a technical foundation has been provided for improving the intelligence of retail scales. However, there is still a technical dilemma of lack of linkage with users' personal smart terminals.

[0007] At the same time, the lack of effective processing methods for a large amount of product data has restricted the development of intelligent product push.

[0008] Chinese patent number CN202110019917.6 discloses a marketing precision screening and push system based on big data and intelligent Internet of Things. The system includes: a mall perception module, a mall operation big data module, a data calculation and analysis module, and a product push module; the mall perception module and the mall operation big data module are used to obtain customer information and sales data inside the mall; the data calculation and analysis module processes customer information and sales data to obtain a store area feature matrix; the product push module analyzes customer information and the store area feature matrix to push products to customers. However, the system lacks mining of product data and obviously has security issues with sensitive data that represents user personal preferences.

[0009] In summary, existing technologies urgently need a technical solution for product push based on AI and deep learning. Summary of the invention

[0010] The purpose of this application is to provide a product push method and device based on AI and deep learning to solve the technical problems raised in the above background technology.

[0011] To achieve the above objectives, this application discloses the following technical solutions:

[0012] In a first aspect, the present application discloses a product push method based on AI and deep learning, the method comprising:

[0013] S1: The first terminal configured with retail scales;

[0014] S2: Pushing products using the first terminal and a preset public push model;

[0015] S3: Acquire a second terminal of an authorized user;

[0016] S4: using the second terminal and a preset personal interaction model to interact with the product;

[0017] The first terminal is used to obtain commodity data and user data, and stores the public push model;

[0018] The public push model is used to store a commodity association database, push existing associated commodities based on the commodity data, generate commodity interaction information based on user data, and update the commodity association database based on one or more of sales records and sales strategies;

[0019] The second terminal is used to obtain the commodity interaction information and stores the personal interaction model;

[0020] The personal interaction model is used to store a personal commodity database, and the personal commodity database is updated based on the commodity interaction information;

[0021] The first terminal and the second terminal are connected in communication after authorization.

[0022] Preferably, the obtaining of product data and user data specifically includes:

[0023] The first terminal is configured using AI technology, wherein the first terminal includes a first camera, a second camera, and corresponding data transmission and storage devices;

[0024] Using the first camera to capture a picture of the product to obtain the corresponding product data, the product data including the product name, product number, product price and corresponding related products, related product prices and related product locations;

[0025] The second camera is used to capture a picture of the user to obtain the authorized user data, where the user data includes user information and protocol information of the second terminal of the corresponding user.

[0026] Preferably, the process of constructing the commodity association database specifically includes:

[0027] The commodity association score is calculated using a commodity association formula to obtain the associated commodity, wherein the commodity association formula is specifically:

[0028]

[0029] Among them, CCS i,j TCS is the product correlation score of products with numbers i and j. 0_ij is the type association score of the preset number i and number j products, D is the preset association reconstruction period, N ij is the number of times that the products with numbers i and j are purchased at the same time in one association reconstruction cycle, CCS j,k TCS is the product association score of products with numbers j and k. 0_jk The type association score for the preset number j and number k products, N jk is the number of times that the products with numbers j and k are purchased at the same time in one association reconstruction cycle, CCS τ is the preset commodity association score threshold, if CCS i,j ≥CCS τ andCCS j,k ≥CCS τ Indicates that when the product association score of the product number i and the product number j is greater than or equal to the product association score threshold and the product association score of the product number j and the product number k is greater than or equal to the product association score threshold, the product number i and the product number k are associated and the product association score of the product number i and the product number k is CCS i,k For CCS i,j and CCS j,k The minimum value in , where min is the minimum value operator.

[0030] Preferably, the pushing of existing related products based on the product data specifically includes:

[0031] The product data is parsed to obtain the corresponding product number and the corresponding associated product, the associated product price and the associated product location, and the associated products arranged from large to small based on the product association score are pushed and displayed on a preset public screen.

[0032] Preferably, the generating of commodity interaction information based on user data specifically includes:

[0033] Based on the user data, the first terminal establishes a communication relationship with the second terminal;

[0034] The personal interaction model is traversed, and when identical commodities exist, the commodity name and the commodity price are obtained based on the commodity data, and the commodity name, the corresponding commodity price and the shopping mall are packaged and defined as the commodity interaction information.

[0035] Preferably, the updating of the commodity association database based on one or more of sales records and sales strategies specifically includes:

[0036] Based on the time series and the association reconstruction period, the commodity association score is recalculated using the commodity association formula, and commodities whose recalculated commodity association scores are greater than or equal to the commodity association score threshold are defined as the associated commodities;

[0037] When the sales strategy exists, the type association score is updated, the product association score is recalculated using the product association formula, and the products whose recalculated product association scores are greater than or equal to the product association score threshold are defined as the associated products.

[0038] Preferably, the process of constructing the public push model specifically includes:

[0039] Use deep learning algorithms to build public push pre-models;

[0040] Training the public push pre-model to acquire product data and user data;

[0041] storing the commodity-related database into the public push pre-model;

[0042] Training the public push pre-model to push existing related products based on the product data;

[0043] Training the public push pre-model to generate commodity interaction information based on user data;

[0044] Training the public push pre-model to execute updating of the commodity association database based on one or more of sales records and sales strategies;

[0045] The commodity association database and the trained public push pre-model are stored and output to obtain the public push model.

[0046] Preferably, the storage includes a personal commodity database, and updating the personal commodity database based on the commodity interaction information specifically includes:

[0047] The personal product database is specifically:

[0048] The second terminal, when authorized, analyzes the user's shopping record and defines the product name and the corresponding product price and shopping mall as the personal product data and stores them in the personal product database;

[0049] The updating of the personal commodity database based on the commodity interaction information is specifically as follows:

[0050] The commodity interaction information is parsed, and based on the commodity name, the commodity interaction information is stored in and updated in the personal commodity database.

[0051] Preferably, the construction process of the personal interaction model includes:

[0052] Use deep learning algorithms to build a personal interaction pre-model;

[0053] storing the personal product database into the personal interaction pre-model;

[0054] Training the personal interaction pre-model to obtain the commodity interaction information and updating the personal commodity database based on the commodity interaction information;

[0055] The personal commodity database and the trained personal interaction pre-model are stored and output to obtain the personal interaction model.

[0056] In a second aspect, the present application discloses a commodity push device based on AI and deep learning, which is applicable to the commodity push method based on AI and deep learning described above, and the device includes:

[0057] A first terminal setting module, wherein the first terminal setting module is configured to: configure a first terminal of a retail scale;

[0058] A commodity push module, wherein the commodity push configuration is: using the first terminal and a preset public push model to push commodities;

[0059] A second terminal setting module, wherein the second terminal setting module is configured to: obtain a second terminal of an authorized user;

[0060] A commodity interaction module, wherein the commodity interaction module is configured to: perform commodity interaction using the second terminal and a preset personal interaction model;

[0061] The first terminal is used to obtain commodity data and user data, and stores the public push model;

[0062] The public push model is used to store a commodity association database, push existing associated commodities based on the commodity data, generate commodity interaction information based on user data, and update the commodity association database based on one or more of sales records and sales strategies;

[0063] The second terminal is used to obtain the commodity interaction information and stores the personal interaction model;

[0064] The personal interaction model is used to store a personal commodity database, and the personal commodity database is updated based on the commodity interaction information;

[0065] The first terminal and the second terminal are connected in communication after authorization.

[0066] Beneficial effects: The product push method and device based on AI and deep learning of the present application solve the problem of displaying push information through a preset public screen, solves the problem of finding products by processing the location of product data, solves the problem of data security by preset protocol information, and uses the first terminal to realize the intelligent improvement of retail scales, improves the technical dilemma of lack of linkage with the user's personal smart terminal, realizes the effective processing of product data through a public push model and a personal interaction model, promotes the development of intelligent product push, and realizes product push based on AI and deep learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0068] Figure 1 A flowchart of a product push method based on AI and deep learning provided in an embodiment of the present application;

[0069] Figure 2 This is a structural block diagram of a product push device based on AI and deep learning provided in an embodiment of the present application. DETAILED DESCRIPTION

[0070] The technical solutions in the embodiments of the present application are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0071] In this article, the term "comprising" is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of more restrictions, the elements defined by the sentence "comprising..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0072] In a first aspect, this embodiment discloses Figure 1 A product push method based on AI and deep learning is shown, the method comprising:

[0073] S1: configuring a first terminal for a retail scale; it should be noted that the retail scale in this embodiment may be, but is not limited to, an existing barcode scale or POS scale, etc.;

[0074] S2: Pushing products using the first terminal and a preset public push model;

[0075] S3: Acquire a second terminal of an authorized user;

[0076] S4: using the second terminal and a preset personal interaction model to interact with the product;

[0077] The first terminal is used to obtain commodity data and user data, and stores a public push model.

[0078] The public push model is used to store a commodity association database, push existing associated commodities based on commodity data, generate commodity interaction information based on user data, and update the commodity association database based on one or more of sales records and sales strategies.

[0079] The second terminal is used to obtain commodity interaction information and stores a personal interaction model.

[0080] The personal interaction model is used to store a personal commodity database and update the personal commodity database based on commodity interaction information.

[0081] The first terminal and the second terminal are connected in communication after authorization.

[0082] Specifically, obtain product data and user data, including:

[0083] The first terminal is configured using AI technology, and the first terminal includes a first camera, a second camera, and corresponding data transmission and storage devices. It should be noted that this embodiment uses existing AI technology to achieve intelligent upgrades to retail scales. For example, by adding an AI module, the module may include but is not limited to a CPU, an NPU, a public push model, a camera, a human body sensor, a storage device, a privacy TPM module, an external plug-in connected to a barcode scale, or connected to a scale via wireless encrypted transmission (WIFI / BT / RF), and synchronously obtains the corresponding PLU and data, namely, product data and user data.

[0084] The first camera is used to capture a picture of the product to obtain corresponding product data, where the product data includes the product name, product number, product price and corresponding related products, related product prices and related product locations.

[0085] The second camera is used to collect the user's picture to obtain the authorized user data, which includes the user information and the protocol information of the corresponding user's second terminal. It should be noted that in this embodiment, the human body is sensed by the human body sensor, and the face recognition is performed by triggering the second camera. In particular, the face recognition of this embodiment is used to distinguish the absolute identity, and only by obtaining the user information, the attempt to interact with the second terminal pointed to by the corresponding user is realized. The data interaction will only be carried out if the user agrees to the authorization and has the protocol information, thereby realizing the push of goods without involving sensitive user information, improving the security of data, and solving the problem of data security in the process of pushing goods.

[0086] Specifically, the process of building a product association database includes:

[0087] The commodity association formula is used to calculate the commodity association score to obtain the associated commodities, wherein the commodity association formula is specifically:

[0088]

[0089] Among them, CCS i,j TCS is the product correlation score of products with numbers i and j. 0_ij is the preset type association score of the products numbered i and j. It should be noted that the type association score is based on the experience known to sales personnel or technical personnel in the field. The method for obtaining the type association score may include, but is not limited to, a score corresponding to the degree of association obtained based on the product type itself and a score corresponding to the degree of association obtained based on a sales strategy, for example, bundling sales, etc. D is the preset association reconstruction period, and N ij is the number of times that products numbered i and j are purchased at the same time in an association reconstruction cycle, CCS j,k TCS is the product association score of products with numbers j and k.0_jk The type association score for the preset number j and number k products, N jk is the number of times that products with numbers j and k are purchased at the same time in an association reconstruction cycle, CCS τ is the preset commodity association score threshold, if CCS i,j ≥CCS τ andCCS j,k ≥CCS τ Indicates that when the product association score of the product number i and the product number j is greater than or equal to the product association score threshold and the product association score of the product number j and the product number k is greater than or equal to the product association score threshold, the product number i and the product number k are associated and the product association score of the product number i and the product number k is CCS i,k For CCS i,j and CCS j,k It should be noted that this embodiment determines whether the commodities are mutually related commodities by determining whether the commodity association score between the commodities is greater than or equal to the commodity association score threshold. Further, the association analysis of at least three commodities is realized through the association between the two commodities, which improves the rationality and objectivity of the commodity association.

[0090] This embodiment realizes the determination of commodity association based on both experience and data drive through the commodity association database, optimizes the processing method of commodity data, and provides a data basis for improving the rationality of commodity push.

[0091] Specifically, push of existing related products based on product data includes:

[0092] The product data is parsed to obtain the corresponding product number and the corresponding related products, related product prices and related product positions, and the related products arranged from large to small based on the product association scores are pushed and displayed on a preset public screen.

[0093] The working principle of the product push in this embodiment is as follows:

[0094] The user places the product on the retail scale with the first terminal, and the retail scale uses the first camera to collect the image of the product, matches the corresponding related products in the product association database, and displays the related products, the prices of the related products and the locations of the related products on the public screen to realize product push. Furthermore, when the user cannot know the location of the product, he can also place the products that may be related to the product on the retail scale through common sense to obtain the location of the product. The public screen is preset and can be, but not limited to, an exemplary barcode scale screen or a large display screen attached to the barcode scale.

[0095] Specifically, the commodity interaction information is generated based on the user data, including:

[0096] Based on the user data, the first terminal establishes a communication relationship with the second terminal. It should be noted that in this embodiment, the communication relationship between the first terminal and the second terminal is established by using the existing communication protocol technology, and there is only information transmission from the first terminal to the second terminal, which greatly protects the user's sensitive information and avoids unnecessary losses or harassment by sales calls caused by the leakage of sensitive information and the leakage of user consumption behavior.

[0097] The personal interaction model is traversed, and the traversal behavior occurs in the second terminal. When the same product exists, the product name and product price are obtained based on the product data, and the product name and the corresponding product price are packaged with the shopping mall to define the product interaction information.

[0098] The working principle of commodity interaction in this embodiment is as follows:

[0099] The first terminal establishes a communication connection with the second terminal using a preset communication protocol, and transmits the commodity data to the second terminal. The personal commodity library matches and updates the personal commodity data in the library based on the commodity data.

[0100] For example, the same product is often sold at A mall, with a price of A1 for 1 catties of bananas, but at B mall, the price is B1 for 1 catties of bananas. However, users often do not remember whether the price is lower in A mall or B mall. When the retail scale is configured with the first terminal, the corresponding price can be synchronized to the user's second terminal, and when the same product is triggered again, the corresponding price will be automatically compared. In other words, in the future, price comparison of products by users will no longer be a pain point, but users can easily check the malls they have visited and find the corresponding product prices. Of course, these data will not exist on the Internet, which means that the data is safe, and only the user's own second data segment has data.

[0101] Specifically, updating the product association database based on one or more of the sales records and the sales strategy specifically includes:

[0102] Based on the time series and the association reconstruction cycle, the product association score is recalculated using the product association formula, and the products whose recalculated product association score is greater than or equal to the product association score threshold are defined as associated products.

[0103] When there is a sales strategy, the type association score is updated, the product association score is recalculated using the product association formula, and the products whose recalculated product association score is greater than or equal to the product association score threshold are defined as associated products.

[0104] In this regard, this embodiment realizes real-time updating of the commodity association database by associating the reconstruction cycle and judging whether there is a sales strategy, thereby optimizing the timeliness of commodity push.

[0105] Specifically, it is characterized in that the construction process of the public push model specifically includes:

[0106] A public push pre-model is constructed using a deep learning algorithm; it should be noted that the construction and training of the model of the deep learning algorithm used here are existing deep learning algorithms and corresponding existing model construction and training technologies;

[0107] Train the public push pre-model to obtain product data and user data;

[0108] Store the product-related database into the public push pre-model;

[0109] Training a public push pre-model to push related products based on product data;

[0110] Train the public push pre-model to generate product interaction information based on user data;

[0111] Training a public push pre-model to execute updating of a product association database based on one or more of sales records and sales strategies;

[0112] The commodity association database and the trained public push pre-model are stored and output to obtain a public push model.

[0113] In this regard, this embodiment achieves optimized product push through a public push model.

[0114] Specifically, a personal product database is stored, and the personal product database is updated based on product interaction information, specifically including:

[0115] Personal product database, specifically:

[0116] The second terminal, when authorized, analyzes the user's shopping record and defines the product name and the corresponding product price and shopping mall as personal product data and stores it in the personal product database;

[0117] Update the personal product database based on product interaction information, specifically:

[0118] Parse the commodity interaction information, store the commodity interaction information in the personal commodity database based on the commodity name, and update the personal commodity database.

[0119] In this regard, this embodiment optimizes the processing of personal commodity data through the personal commodity database.

[0120] Specifically, the construction process of the personal interaction model includes:

[0121] A personal interaction pre-model is constructed using a deep learning algorithm; it should be noted that the construction and training of the model of the deep learning algorithm used here is an existing deep learning algorithm and the corresponding existing model construction and training technology;

[0122] Store the personal product database into the personal interaction pre-model;

[0123] Training the personal interaction pre-model to obtain product interaction information and update the personal product database based on the product interaction information;

[0124] The personal product database and the trained personal interaction pre-model are output to obtain a personal interaction model.

[0125] In this regard, this embodiment optimizes the processing of the user's personal shopping big data through a personal interaction model, providing an accurate data basis for the user to select goods, and the model only exists in the user's second terminal, ensuring the security of the user's sensitive data. In this embodiment, the second terminal can be but is not limited to a smart phone, etc.

[0126] In a second aspect, this embodiment discloses Figure 2 A commodity push device based on AI and deep learning is shown, and the device is applicable to a commodity push method based on AI and deep learning, and the device includes:

[0127] A first terminal setting module, the first terminal setting module is configured to: configure a first terminal of a retail scale;

[0128] The product push module is configured to push the product using the first terminal and a preset public push model;

[0129] A second terminal setting module, the second terminal setting module is configured to: obtain a second terminal of an authorized user;

[0130] The commodity interaction module is configured to: use the second terminal and a preset personal interaction model to perform commodity interaction.

[0131] The first terminal is used to obtain commodity data and user data, and stores a public push model.

[0132] The public push model is used to store a commodity association database, push existing associated commodities based on commodity data, generate commodity interaction information based on user data, and update the commodity association database based on one or more of sales records and sales strategies.

[0133] The second terminal is used to obtain commodity interaction information and stores a personal interaction model.

[0134] The personal interaction model is used to store a personal commodity database and update the personal commodity database based on commodity interaction information.

[0135] The first terminal and the second terminal are connected in communication after authorization.

[0136] It should be noted that the product push device based on AI and deep learning in this embodiment corresponds to the product push method based on AI and deep learning in the previous text. Therefore, the content not specifically described in the product push device based on AI and deep learning in this embodiment may have but is not limited to the same functional definition, working principle and technical effect as the product push method based on AI and deep learning in the previous text, and this text will not go into details here.

[0137] In summary, the product push method and device based on AI and deep learning in this embodiment solve the problem of displaying push information through a preset public screen, solves the problem of finding products by processing the location of product data, solves the problem of data security by preset protocol information, and uses the first terminal to achieve intelligent improvement of the retail scale, improves the technical dilemma of lack of linkage with the user's personal smart terminal, realizes effective processing of product data through a public push model and a personal interaction model, promotes the development of intelligent product push, and realizes product push based on AI and deep learning.

[0138] In the embodiments provided in the present application, it should be understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code or any appropriate combination thereof. For hardware implementation, the processor can be implemented in one or more of the following units: application specific integrated circuit (ASIC), digital signal processor (DSP), digital signal processing device (DSPD), programmable logic device (PLD), field programmable gate array (FPGA), processor, controller, microcontroller, microprocessor, other electronic units designed to implement the functions described herein or their combination. For software implementation, part or all of the flow of the embodiment can be completed by instructing the relevant hardware through a computer program. When implemented, the above program can be stored in a computer-readable storage medium or transmitted as one or more instructions or codes on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein the communication medium includes any medium that is convenient for transmitting a computer program from one place to another. The storage medium can be any available medium that a computer can access. The computer-readable storage medium can include but is not limited to RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer.

[0139] Finally, it should be noted that the above is only a preferred embodiment of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A product push method based on AI and deep learning, characterized in that: The method includes: S1: The first terminal configured with retail scales; S2: Pushing products using the first terminal and a preset public push model; S3: Acquire a second terminal of an authorized user; S4: using the second terminal and a preset personal interaction model to interact with the product; The first terminal is used to obtain commodity data and user data, and stores the public push model; The public push model is used to store a commodity association database, push existing associated commodities based on the commodity data, generate commodity interaction information based on user data, and update the commodity association database based on one or more of sales records and sales strategies; The second terminal is used to obtain the commodity interaction information and stores the personal interaction model; The personal interaction model is used to store a personal commodity database, and the personal commodity database is updated based on the commodity interaction information; The first terminal and the second terminal are connected in communication after authorization.

2. The method for pushing products based on AI and deep learning according to claim 1, characterized in that: The obtaining of product data and user data specifically includes: The first terminal is configured using AI technology, wherein the first terminal includes a first camera, a second camera, and corresponding data transmission and storage devices; Using the first camera to capture a picture of the product to obtain the corresponding product data, the product data including the product name, product number, product price and corresponding related products, related product prices and related product locations; The second camera is used to capture a picture of the user to obtain the authorized user data, where the user data includes user information and protocol information of the second terminal of the corresponding user.

3. The method for pushing products based on AI and deep learning according to claim 2, characterized in that: The process of constructing the commodity association database specifically includes: The commodity association score is calculated using a commodity association formula to obtain the associated commodity, wherein the commodity association formula is specifically: Among them, CCS i,j TCS is the product correlation score of products with numbers i and j. 0_ij is the type association score of the preset number i and number j products, D is the preset association reconstruction period, N ij is the number of times that the products with numbers i and j are purchased at the same time in one association reconstruction cycle, CCS j,k TCS is the product association score of products with numbers j and k. 0_jk The type association score for the preset number j and number k products, N jk is the number of times that the products with numbers j and k are purchased at the same time in one association reconstruction cycle, CCS τ is the preset commodity association score threshold, if CCS i,j ≥CCS τ andCCS j,k ≥CCS τ Indicates that when the product association score of the product number i and the product number j is greater than or equal to the product association score threshold and the product association score of the product number j and the product number k is greater than or equal to the product association score threshold, the product number i and the product number k are associated and the product association score of the product number i and the product number k is CCS i,k For CCS i,j and CCS j,k The minimum value in , where min is the minimum value operator.

4. The method for pushing products based on AI and deep learning according to claim 3, characterized in that: The pushing of existing related products based on the product data specifically includes: The product data is parsed to obtain the corresponding product number and the corresponding associated product, the associated product price and the associated product location, and the associated products arranged from large to small based on the product association score are pushed and displayed on a preset public screen.

5. The method for pushing products based on AI and deep learning according to claim 2, characterized in that: The generating of commodity interaction information based on user data specifically includes: Based on the user data, the first terminal establishes a communication relationship with the second terminal; The personal interaction model is traversed, and when identical commodities exist, the commodity name and the commodity price are obtained based on the commodity data, and the commodity name, the corresponding commodity price and the shopping mall are packaged and defined as the commodity interaction information.

6. The method for pushing products based on AI and deep learning according to claim 3, characterized in that: The updating of the commodity association database based on one or more of the sales records and the sales strategies specifically includes: Based on the time series and the association reconstruction period, the commodity association score is recalculated using the commodity association formula, and commodities whose recalculated commodity association scores are greater than or equal to the commodity association score threshold are defined as the associated commodities; When the sales strategy exists, the type association score is updated, the product association score is recalculated using the product association formula, and the products whose recalculated product association scores are greater than or equal to the product association score threshold are defined as the associated products.

7. The method for pushing products based on AI and deep learning according to any one of claims 2 to 6, characterized in that: The construction process of the public push model specifically includes: Use deep learning algorithms to build public push pre-models; Training the public push pre-model to acquire product data and user data; storing the commodity-related database into the public push pre-model; Training the public push pre-model to push existing related products based on the product data; Training the public push pre-model to generate commodity interaction information based on user data; Training the public push pre-model to execute updating of the commodity association database based on one or more of sales records and sales strategies; The commodity association database and the trained public push pre-model are stored and output to obtain the public push model.

8. The method for pushing products based on AI and deep learning according to claim 5, characterized in that: The storage includes a personal commodity database, and updating the personal commodity database based on the commodity interaction information specifically includes: The personal product database is specifically: The second terminal, when authorized, analyzes the user's shopping record and defines the product name and the corresponding product price and shopping mall as the personal product data and stores them in the personal product database; The updating of the personal commodity database based on the commodity interaction information is specifically as follows: The commodity interaction information is parsed, and based on the commodity name, the commodity interaction information is stored in and updated in the personal commodity database.

9. The method for pushing products based on AI and deep learning according to claim 8, characterized in that: The construction process of the personal interaction model includes: Use deep learning algorithms to build a personal interaction pre-model; storing the personal product database into the personal interaction pre-model; Training the personal interaction pre-model to obtain the commodity interaction information and updating the personal commodity database based on the commodity interaction information; The personal commodity database and the trained personal interaction pre-model are stored and output to obtain the personal interaction model.

10. A commodity push device based on AI and deep learning, the device is applicable to the commodity push method based on AI and deep learning as described in any one of claims 1-9, characterized in that: The device includes: A first terminal setting module, wherein the first terminal setting module is configured to: configure a first terminal of a retail scale; A commodity push module, wherein the commodity push configuration is: using the first terminal and a preset public push model to push commodities; A second terminal setting module, wherein the second terminal setting module is configured to: obtain a second terminal of an authorized user; A commodity interaction module, wherein the commodity interaction module is configured to: perform commodity interaction using the second terminal and a preset personal interaction model; The first terminal is used to obtain commodity data and user data, and stores the public push model; The public push model is used to store a commodity association database, push existing associated commodities based on the commodity data, generate commodity interaction information based on user data, and update the commodity association database based on one or more of sales records and sales strategies; The second terminal is used to obtain the commodity interaction information and stores the personal interaction model; The personal interaction model is used to store a personal commodity database, and the personal commodity database is updated based on the commodity interaction information; The first terminal and the second terminal are connected in communication after authorization.

Images