Product recommendation method, device and equipment based on user facial expression

Abstract

The invention relates to a product recommendation method, device and equipment based on user facial expressions. The method includes: acquiring facial expression information of a to-be-recommended user; calculating an actual micro-expression of the user to be recommended according to the facial expression information; determining a standard micro-expression corresponding to the actual micro-expression according to a pre-trained difference feature model; determining a recommended commodity set corresponding to the standard micro-expression according to a pre-trained standard micro-expression-commodity set mapping library of the to-be-recommended user; and recommending the commodities in the recommended commodity set to the to-be-recommended user. According to the embodiment of the invention, the method adapts to the difference between facial expressions of different users, and improves the universality and accuracy of product recommendation.

Description

Technical Field

[0001] The embodiments of this specification relate to the field of artificial intelligence technology, and more particularly to a method, device and apparatus for recommending products based on user facial expressions. Background Art

[0002] With the development of financial technology, the service model has begun to change from the traditional offline face-to-face method to the remote screen-to-screen method. Insights into customers' true intentions and preferences through their facial expressions can help banks improve their service capabilities. However, since everyone has their own unique faces, the facial expressions of different users vary greatly, and the current product recommendation methods have poor universality and low accuracy.

[0003] How to improve the universality and accuracy of product recommendations is a technical problem that needs to be solved urgently. Summary of the invention

[0004] In order to solve the problems of poor universality and low accuracy of product recommendation methods in the prior art, the embodiments of this specification provide a product recommendation method, device and equipment based on user facial expressions, train a standard micro-expression-product set mapping library of each user and a difference feature model between the user's actual micro-expressions and the standard micro-expressions, determine the standard micro-expressions corresponding to the actual micro-expressions of the user to be recommended according to the difference feature model of the user to be recommended, determine the product set corresponding to the standard micro-expression from the standard micro-expression-product set mapping library, and recommend the products in the product set to the user to be recommended.

[0005] The specific technical solutions of the embodiments of this specification are as follows:

[0006] On the one hand, the embodiments of this specification provide a product recommendation method based on user facial expressions, including:

[0007] Obtain facial expression information of the user to be recommended;

[0008] Calculating the actual micro-expression of the user to be recommended according to the facial expression information;

[0009] Determine a standard micro-expression corresponding to the actual micro-expression according to a pre-trained difference feature model;

[0010] Determine the recommended product set corresponding to the standard micro-expression according to the pre-trained standard micro-expression-product set mapping library of the user to be recommended;

[0011] Recommend the products in the recommended product set to the user to be recommended.

[0012] Furthermore, the step of training the difference feature model includes:

[0013] Acquire multiple facial expression information of multiple training users and respectively calculate multiple first training micro-expressions corresponding to the multiple facial expression information of each training user and multiple emotion information corresponding to the multiple facial expression information;

[0014] Determine at least one standard micro-expression corresponding to each emotional information of each training user according to a pre-constructed standard micro-expression-emotion library, wherein the standard micro-expression-emotion library includes a correspondence between a plurality of standard micro-expressions and emotional information;

[0015] Calculate the difference between each standard micro-expression and the corresponding first training micro-expression;

[0016] The difference feature model is constructed according to the multiple differences of each training user.

[0017] Furthermore, respectively calculating the difference between each standard micro-expression and the corresponding first training micro-expression further includes:

[0018] Vectorizing the first training micro-expression to obtain a first feature vector;

[0019] Vectorizing the standard micro-expression to obtain a second feature vector;

[0020] The difference feature between the first feature vector and the second feature vector is calculated to obtain the difference between the corresponding standard micro-expression and the first training micro-expression.

[0021] Furthermore, the formula for calculating the difference feature between the first feature vector and the second feature vector is:

[0022]

[0023] Wherein, DF represents the difference feature, DF1 represents the first feature vector, and DF2 represents the second feature vector.

[0024] Furthermore, constructing the difference feature model according to the multiple differences of each training user further includes:

[0025] Calculating a linear relationship between a first training micro-expression and a standard micro-expression of each training user according to a plurality of difference feature vectors of each training user;

[0026] The linear relationship is used as the difference feature model.

[0027] Furthermore, training the to-be-recommended user standard micro-expression-product set mapping library further includes:

[0028] Constructing an initial standard micro-expression-product set mapping library for the user to be recommended;

[0029] Acquire facial expression information of the user to be recommended, and calculate a second training micro-expression of the user to be recommended;

[0030] Determine, according to a pre-trained difference feature model, a standard micro-expression corresponding to the second training micro-expression of the user to be recommended;

[0031] Determine a predicted recommended commodity set corresponding to the standard micro-expression of the user to be recommended according to the initial standard micro-expression-commodity set mapping library, and recommend the predicted recommended commodity set to the user to be recommended;

[0032] Acquiring operation information of the to-be-recommended user on the predicted recommended product set, and calculating a trend intention evaluation value according to the operation information;

[0033] Determining whether the trend intention evaluation value exceeds a first threshold;

[0034] If not, adjusting the correspondence between the standard micro-expression and the commodity set in the initial standard micro-expression-commodity set mapping library of the user to be recommended, and repeating the steps of obtaining the facial expression information of the second training user and calculating the second training micro-expression of the second training user;

[0035] If so, a trained standard micro-expression-product set mapping library is obtained.

[0036] Furthermore, the operation information at least includes:

[0037] Product interest behavior data and product transaction conversion data.

[0038] Furthermore, calculating the trend intention evaluation value according to the operation information further includes:

[0039] Calculating the commodity interest behavior data to obtain a first valuation;

[0040] Calculating the commodity transaction conversion data to obtain a second estimated value;

[0041] Performing weighted calculation on the first estimate and the second estimate to obtain a data plane estimate;

[0042] Based on the commodity interest behavior data and commodity transaction conversion data, clustering the predicted recommended commodity set using a clustering algorithm, and calculating the information surface valuation according to the clustering result;

[0043] Calculating the knowledge valuation through knowledge induction according to the regular preference of the to-be-recommended user for each commodity in the predicted recommended commodity set;

[0044] The data aspect valuation, information aspect valuation and knowledge aspect valuation are weightedly calculated to obtain the trend intention evaluation value.

[0045] Furthermore, the product interest behavior data includes at least the number of product clicks and the product browsing time.

[0046] Furthermore, calculating the commodity interest behavior data to obtain a first valuation further includes:

[0047] The first estimated value is obtained by performing a weighted calculation on the number of clicks on the product and the browsing time of the product.

[0048] Furthermore, the commodity transaction conversion data at least includes commodity transaction volume, secondary transaction volume and number of negative reviews.

[0049] Furthermore, calculating the commodity transaction conversion data to obtain a second estimated value further includes:

[0050] Calculating the sum of the commodity transaction volume and the secondary transaction volume to obtain the total commodity transaction volume;

[0051] The ratio of the total commodity transaction volume to the number of negative evaluations is calculated to obtain the second estimated value.

[0052] Furthermore, if the trend intention evaluation value exceeds the first threshold, the method further includes:

[0053] Determining whether the trend intention evaluation value exceeds a second threshold;

[0054] If so, the standard micro-expression corresponding to the predicted recommended product set in the initial standard micro-expression-product set mapping library is replaced with the second training micro-expression.

[0055] Furthermore, if the trend intention evaluation value exceeds the second threshold, the method further includes:

[0056] taking a second training micro-expression corresponding to the trend intention evaluation value exceeding the second threshold as a reference micro-expression;

[0057] Performing weighted calculation on multiple reference micro-expressions corresponding to the same standard micro-expression according to the corresponding trend intention evaluation values ​​to obtain an updated micro-expression;

[0058] The standard micro-expression is replaced with the updated micro-expression.

[0059] Furthermore, after determining the recommended product set corresponding to the standard micro-expression according to the pre-trained standard micro-expression-product set mapping library of the user to be recommended, the method further includes:

[0060] Eliminate products that do not match the user information of the user to be recommended from the recommended product set;

[0061] Recommending the products in the recommended product set to the to-be-recommended user further includes:

[0062] The remaining commodities in the recommended commodity set are recommended to the user to be recommended.

[0063] On the other hand, the embodiment of this specification also provides a product recommendation device based on user facial expressions, the device comprising:

[0064] A facial expression information acquisition unit, used to acquire facial expression information of the user to be recommended;

[0065] An actual micro-expression calculation unit, used to calculate the actual micro-expression of the user to be recommended according to the facial expression information;

[0066] A standard micro-expression determination unit, used to determine a standard micro-expression corresponding to the actual micro-expression according to a pre-trained difference feature model;

[0067] A recommended product set determination unit, configured to determine a recommended product set corresponding to the standard micro-expression according to a pre-trained standard micro-expression-product set mapping library of the user to be recommended;

[0068] The commodity recommendation unit is used to recommend the commodities in the recommended commodity set to the user to be recommended.

[0069] On the other hand, an embodiment of the present specification further provides a computer device, including a memory, a processor, and a computer program stored in the memory, and the processor implements the above method when executing the computer program.

[0070] On the other hand, an embodiment of the present specification further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program implements the above method when executed by a processor.

[0071] Finally, the embodiments of this specification also provide a computer program product, wherein the computer program product includes a computer program, and the computer program implements the above method when executed by a processor.

[0072] By using the embodiments of the present specification, a standard micro-expression-product set mapping library of each user and a difference feature model between the user's actual micro-expressions and the standard micro-expressions are trained, and the standard micro-expressions corresponding to the actual micro-expressions of the user to be recommended are determined based on the difference feature model of the user to be recommended, and the product set corresponding to the standard micro-expression is determined from the standard micro-expression-product set mapping library, and the products in the product set are recommended to the user to be recommended, thereby adapting to the differences between the facial expressions of different users and improving the universality and accuracy of product recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] In order to more clearly illustrate the embodiments of this specification 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 embodiments of this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0074] Figure 1 It is a schematic diagram of a system for implementing a product recommendation method based on user facial expressions in an embodiment of this specification;

[0075] Figure 2 The figure is a flow chart of a shared access method based on a cloud-native storage data volume according to an embodiment of the present specification;

[0076] Figure 3 The figure is a schematic diagram of the process of training the difference feature model in the embodiment of this specification;

[0077] Figure 4 It is a schematic diagram of a process of respectively calculating the difference between each standard micro-expression and the corresponding first training micro-expression in the embodiment of this specification;

[0078] Figure 5 It is a schematic diagram of a process of constructing the difference feature model according to multiple differences of each training user in the embodiment of this specification;

[0079] Figure 6 The figure is a schematic diagram of the process of training the standard micro-expression-product set mapping library of the user to be recommended in the embodiment of this specification;

[0080] Figure 7 It is a schematic diagram of a process of calculating a trend intention evaluation value according to the operation information in an embodiment of this specification;

[0081] Figure 8 The figure shows a flow chart of calculating the commodity transaction conversion data to obtain a second estimated value in the embodiment of this specification;

[0082] Fig. 9 It is a schematic diagram of the process of updating the standard micro-expression in the embodiment of this specification;

[0083] Fig.10 It is a schematic diagram of the process of updating the standard micro-expression in another embodiment of the present specification;

[0084] Fig.11 The figure shows a flow chart of removing commodities that do not match the user information of the user to be recommended in the embodiment of the present specification;

[0085] Fig.12The figure shows a schematic diagram of the structure of a product recommendation device based on user facial expressions in an embodiment of this specification;

[0086] Fig.13 The figure is a schematic diagram of the structure of a computer device in an embodiment of the present specification.

[0087]

Description of reference numerals

[0088] 101. Terminal;

[0089] 102. Server;

[0090] 1201, facial expression information acquisition unit;

[0091] 1202. Actual micro-expression calculation unit;

[0092] 1203, standard micro-expression determination unit;

[0093] 1204, a recommended product set determination unit;

[0094] 1205. Product recommendation unit;

[0095] 1302. Computer equipment;

[0096] 1304. Processing equipment;

[0097] 1306. Storage resources;

[0098] 1308. Drive system;

[0099] 1310, input / output module;

[0100] 1312. Input devices;

[0101] 1314. Output device;

[0102] 1316. Presentation equipment;

[0103] 1318. Graphical user interface;

[0104] 1320, network interface;

[0105] 1322, communication link;

[0106] 1324. Communication bus. DETAILED DESCRIPTION

[0107] The following will be combined with the drawings in the embodiments of this specification to clearly and completely describe the technical solutions in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. Based on the embodiments in the embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the embodiments of this specification.

[0108] It should be noted that the terms "first", "second", etc. in the description and claims of the embodiments of this specification and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the embodiments of this specification described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, device, product or equipment that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0109] It should be noted that the acquisition, storage, use, and processing of data in the technical solutions of the embodiments of this specification comply with the relevant provisions of national laws and regulations.

[0110] It should be noted that in the embodiments of this specification, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of this application, but it does not mean that the applicant has or will necessarily use the solution.

[0111] like Figure 1 The diagram is a schematic diagram of a system for implementing a product recommendation method based on user facial expressions in an embodiment of the present specification, including a terminal 101 and a server 102. The terminal 101 and the server 102 can communicate with each other through a network, and the network can include a local area network (LAN), a wide area network (WAN), the Internet, or a combination thereof, and is connected to a website, a user device (such as a computing device), and a backend system.

[0112] Terminal 101 collects the user's facial expression information and sends the facial expression information to server 102. Server 102 calculates the facial expression information using a stored pre-trained difference feature model and a standard micro-expression-product set mapping library to obtain a recommended product set, and recommends the products in the recommended product set to the user through terminal 101.

[0113] Alternatively, the server 102 may be a node of a cloud computing system (not shown), or each server may be a separate cloud computing system including a plurality of computers interconnected by a network and operating as a distributed processing system.

[0114] In addition, it should be noted that Figure 1 What is shown is only one application environment provided by the embodiment of this specification. In actual application, other application environments may also be included, and this specification does not limit it.

[0115] In response to the problems existing in the prior art, the embodiments of this specification provide a product recommendation method based on user facial expressions, which trains a standard micro-expression-product set mapping library for each user and a difference feature model between the user's actual micro-expressions and the standard micro-expressions, determines the standard micro-expressions corresponding to the actual micro-expressions of the user to be recommended based on the difference feature model of the user to be recommended, determines the product set corresponding to the standard micro-expressions from the standard micro-expressions-product set mapping library, and recommends the products in the product set to the user to be recommended. Figure 2 The flowchart of the product recommendation method based on the user's facial expression in the embodiment of this specification is shown. This figure describes the process of recommending products to users based on their facial expressions. The order of steps listed in the embodiment is only one way of executing the steps among many orders, and does not represent the only order of execution. When the system or device product is executed in practice, it can be executed in the order of the method shown in the embodiment or the accompanying drawings or in parallel.

[0116] Specific as Figure 2 As shown, the method may include:

[0117] Step 201: Obtain facial expression information of the user to be recommended;

[0118] Step 202: Calculating the actual micro-expression of the user to be recommended according to the facial expression information;

[0119] Step 203: determining the standard micro-expression corresponding to the actual micro-expression according to the pre-trained difference feature model;

[0120] Step 204: determining a recommended product set corresponding to the standard micro-expression according to a pre-trained standard micro-expression-product set mapping library of the user to be recommended;

[0121] Step 205: Recommend the products in the recommended product set to the user to be recommended.

[0122] By using the embodiments of the present specification, a standard micro-expression-product set mapping library of each user and a difference feature model between the user's actual micro-expressions and the standard micro-expressions are trained, and the standard micro-expressions corresponding to the actual micro-expressions of the user to be recommended are determined based on the difference feature model of the user to be recommended, and the product set corresponding to the standard micro-expression is determined from the standard micro-expression-product set mapping library, and the products in the product set are recommended to the user to be recommended, thereby adapting to the differences between the facial expressions of different users and improving the universality and accuracy of product recommendations.

[0123] In the embodiments of the present specification, facial expression information may be information in the form of video. When a user browses products through a mobile banking app, the mobile app obtains the user's facial video through the mobile phone's camera after obtaining the user's authorization. The mobile app then sends the facial video to a bank server. The bank server calculates the user's actual micro-expressions based on the facial video. The actual micro-expressions may be in the form of vectorized data, that is, the facial video is vectorized to obtain the actual micro-expressions.

[0124] Because facial expressions are different for different people, the actual micro-expressions obtained by quantization of multiple users also vary greatly. In order to recommend products to all users based on their actual micro-expressions, the embodiment of this specification determines the standard micro-expressions corresponding to the actual micro-expressions through a pre-trained difference feature model. The difference feature recognition model can determine the standard micro-expressions corresponding to the actual micro-expressions of all users.

[0125] Then, the recommended product set corresponding to the standard micro-expression is determined based on the pre-trained standard micro-expression-product set mapping library of the user to be recommended. Each user has his own standard micro-expression-product set mapping library. The standard micro-expression-product set mapping library includes the correspondence between multiple standard micro-expressions and multiple product sets. The product set may include different types of product sets or product sets of the same type but different levels, such as financial product sets, fund product sets, current financial product sets, fixed-term financial product sets, etc., which are not elaborated in the embodiments of this specification.

[0126] The standard micro-expressions can form a standard micro-expression library, which is applicable to all users. The trained difference feature model is used to determine the standard micro-expressions corresponding to the actual micro-expressions of each user.

[0127] Finally, the products in the recommended product set are recommended to the user. For example, when the user browses products through the mobile banking app, the next product is a product in the recommended product set.

[0128] According to one embodiment of this specification, Figure 3 As shown, the steps of training the difference feature model include:

[0129] Step 301: Acquire multiple facial expression information of multiple training users and respectively calculate multiple first training micro-expressions corresponding to the multiple facial expression information of each training user and multiple emotion information corresponding to the multiple facial expression information;

[0130] Step 302: determining at least one standard micro-expression corresponding to each emotional information of each training user according to a pre-constructed standard micro-expression-emotion library, wherein the standard micro-expression-emotion library includes a correspondence between a plurality of standard micro-expressions and emotional information;

[0131] Step 303: Calculate the difference between each standard micro-expression and the corresponding first training micro-expression;

[0132] Step 304: construct the difference feature model according to the multiple differences of each training user.

[0133] In the embodiments of the present specification, the user's facial expression information can be recognized by existing emotion recognition methods to obtain corresponding emotion information, which may be happiness or anger, etc.

[0134] The pre-constructed standard micro-expression-emotion library includes the correspondence between multiple standard micro-expressions and emotions. The corresponding emotional information of multiple standard micro-expressions can be divided in advance based on expert experience. One emotional information can correspond to multiple standard micro-expressions. Finally, these correspondences are stored in the standard micro-expression-emotion library.

[0135] In the embodiment of the present specification, after determining the emotion information corresponding to the facial expression information of the training user, a plurality of standard micro-expressions corresponding to the emotion information are determined from a standard micro-expression-emotion library.

[0136] The difference between each standard micro-expression and the first training micro-expression is calculated respectively, and a difference feature model is constructed based on the difference.

[0137] In the embodiments of this specification, Figure 4 As shown, respectively calculating the difference between each standard micro-expression and the corresponding first training micro-expression further includes:

[0138] Step 401: vectorizing the first training micro-expression to obtain a first feature vector;

[0139] Step 402: vectorizing the standard micro-expression to obtain a second feature vector;

[0140] Step 403: Calculate the difference feature between the first feature vector and the second feature vector to obtain the difference between the corresponding standard micro-expression and the first training micro-expression.

[0141] In an embodiment of the present specification, the first training micro-expression can be vectorized by optical flow, histogram of directional optical flow (HOOF) and main direction mean optical flow (MDMO) to obtain a micro-expression feature set, and the first feature vector corresponding to the micro-expression can be obtained by feature fusion.

[0142] For example:

[0143] The optical flow method uses the characteristics of the image sequence in the time domain to calculate the correlation between the pixels in the current frame and the target frame to express the correspondence between the frames, which is used to represent the relationship between the frames of the micro-expression sequence; for the pixel value F(x) at point x in the kth frame and the pixel value G(x) at point x in the k+1th frame, there is a certain disparity h:

[0144] F(x+h)=G(x)

[0145]

[0146] Repeating this process to obtain a series of Newton iterations, h will converge to an optimal value, where the iteration h (k+1) It can be expressed as:

[0147] h0=0

[0148]

[0149] Among them, h k represents the disparity of the kth frame, f′(x+h k ) can be expressed as:

[0150]

[0151] Among them, function F represents the pixel value.

[0152] Then the optical flow feature UIMES_21 from the 1st frame to the Kth frame can be expressed as:

[0153]

[0154] in, represents the disparity of the xth point in the kth frame.

[0155] The Histogram of Oriented Optical Flow (HOOF) eliminates the influence of horizontal motion on optical flow by redefining the direction of optical flow and eliminates the influence of lens distance on optical flow extraction; for the optical flow vector h of the pixel value F(x) of point x in the kth frame and the pixel value G(x) of point x in the k+1th frame, the HOOF is calculated by grading the optical flow vector h based on the direction and weighting it by the size of the vector:

[0156] h=[x,y] T

[0157] Wherein, x and y represent the horizontal and vertical coordinates of the pixel in the kth frame respectively.

[0158] Then calculate the angle θ between the optical flow vector and the horizontal axis:

[0159] θ=tan -1 (y / x)

[0160] Bin the angle θ. When the angle θ falls to:

[0161]

[0162] Wherein, B represents the total number of bins, and b represents that the angle θ falls into the bth bin.

[0163] Its amplitude Act on the bth bin of the histogram, where 1≤b≤B, and finally normalize the histogram. The optical flow vector histogram of each frame at time t can be expressed as:

[0164] UIMES_22=[h t；1 ,h t；2 ,h t；3 ,…,h t；B ] T

[0165] The main direction mean optical flow (MDMO) divides the face into 36 regions of interest and extracts the main direction mean optical flow features of each region from the face image sequence, where the main direction optical flow average value of the i region in the kth frame is Can be identified as:

[0166]

[0167] Among them, B max is the set with the largest number of direction vectors in the statistical histogram, among which, Represents the directional optical flow histogram at position p.

[0168]

[0169] It should be noted that the optical flow method (Optical flow), the histogram of directional optical flow (HOOF) and the main direction mean optical flow (MDMO) are all common knowledge in the art, and the embodiments of this specification are only used for exemplary purposes.

[0170] After principal component analysis (PCA) of the features in the micro-expression feature set UIMES_2 = {UIMES_21, UIMES_22, UIMES_23}, the first feature vector DF1 is obtained. Among them, UIMES_21, UIMES_22 and UIMES_23 are the features obtained by optical flow method, directional optical flow histogram and main direction average optical flow respectively:

[0171] DF1=PCA(UIMES_2)

[0172] The standard micro-expression is vectorized using the same vectorization method to obtain the second feature vector DF2.

[0173] Then calculate the difference between the first and second eigenvectors:

[0174]

[0175] Wherein, DF represents the difference feature, DF1 represents the first feature vector, and DF2 represents the second feature vector.

[0176] Because each emotional information in the standard micro-expression-emotion database corresponds to multiple standard micro-expressions, and one training user corresponds to multiple first training micro-expressions, the difference feature vector DF between the first feature vector DF1 of the first training micro-expression and the second feature vector DF2 of multiple standard micro-expressions will be calculated to obtain multiple difference feature vectors DF. Therefore, multiple difference feature vectors DF of multiple training users under one emotional information will be calculated to form a difference feature vector set F i , where i represents the i-th emotional information. The set of difference feature vectors corresponding to all emotions in the standard micro-expression-emotion library is F = [F1, F2, F3, ..., F n ], where n represents the total number of sexy messages.

[0177] According to one embodiment of this specification, Figure 5 As shown, constructing the difference feature model according to the multiple differences of each training user further includes:

[0178] Step 501: Calculate the linear relationship between the first training micro-expression and the standard micro-expression of each training user according to a plurality of difference feature vectors of each training user;

[0179] Step 502: Use the linear relationship as the difference feature model.

[0180] In the embodiments of the present specification, there are differences between the actual micro-expressions of the user and the standard micro-expressions, and all the differences (that is, the difference feature vectors DF between each standard micro-expression under all emotional information and the first training micro-expressions of multiple users) can be calculated. Therefore, the linear relationship between the first training micro-expressions and the standard micro-expressions can be fitted according to the difference feature vectors, and the linear relationship can be used as the difference feature model.

[0181] When recommending products to the user to be recommended, the linear relationship is used to calculate the standard micro-expression corresponding to the actual micro-expression of the user to be recommended.

[0182] According to one embodiment of this specification, Figure 6 As shown, training the to-be-recommended user standard micro-expression-product set mapping library further includes:

[0183] Step 601: constructing an initial standard micro-expression-product set mapping library of the user to be recommended;

[0184] Step 602: Acquire facial expression information of the user to be recommended, and calculate a second training micro-expression of the user to be recommended;

[0185] Step 603: determining the standard micro-expression corresponding to the second training micro-expression of the user to be recommended according to the pre-trained difference feature model;

[0186] Step 604: determining a predicted recommended commodity set corresponding to the standard micro-expression of the user to be recommended according to the initial standard micro-expression-commodity set mapping library, and recommending the predicted recommended commodity set to the user to be recommended;

[0187] Step 605: obtaining the operation information of the to-be-recommended user on the predicted recommended product set, and calculating the trend intention evaluation value according to the operation information;

[0188] Step 606: Determine whether the trend intention evaluation value exceeds a first threshold;

[0189] Step 607: If not, adjusting the correspondence between the standard micro-expression and the commodity set in the initial standard micro-expression-commodity set mapping library of the user to be recommended, and repeating the steps of obtaining the facial expression information of the second training user and calculating the second training micro-expression of the second training user;

[0190] Step 608: If yes, then a trained standard micro-expression-product set mapping library is obtained.

[0191] In the embodiment of the present specification, the initial standard micro-expression-commodity set mapping library can be constructed based on experience, and the initial standard micro-expression-commodity set mapping library of all users can be the same.

[0192] Then use the embodiment of this manual Figure 5 The method is used to determine the standard micro-expression corresponding to the user's second training micro-expression, and then the predicted recommended product set corresponding to the standard micro-expression is determined according to the initial standard micro-expression-product set mapping library.

[0193] Because the predicted recommended product set is not the most suitable for the user's needs, after the products in the predicted recommended product set are recommended to the user, the user's operation information on the products in the predicted recommended product set is obtained, and the trend intention evaluation value is calculated based on the operation information.

[0194] Then, it is determined whether the trend intention evaluation value exceeds a first threshold value. If so, it indicates that the predicted recommended product set meets the user's needs, and the user's standard micro-expression-product set mapping training is completed.

[0195] If not, it means that the predicted recommended product set does not meet the user's needs. It is necessary to adjust the correspondence between the standard micro-expressions and the product set in the user's standard micro-expression-product set mapping library, and repeatedly obtain the user's facial expression information for iterative training until the tendency intention evaluation values ​​of the second training micro-expressions under all emotional information exceed the first threshold, and a trained standard micro-expression-product set mapping library is obtained.

[0196] Exemplarily, the correspondence between the standard micro-expressions and the commodity sets in the standard micro-expression-commodity set mapping library may be replaced in sequence, which is not limited in the embodiments of this specification.

[0197] It should be noted that the first threshold can be set based on experience, which will not be elaborated in the embodiments of this specification.

[0198] In the embodiment of the present specification, the operation information includes at least: commodity interest behavior data and commodity transaction conversion data.

[0199] like Figure 7 As shown, calculating the trend intention evaluation value according to the operation information further includes:

[0200] Step 701: Calculate the commodity interest behavior data to obtain a first valuation;

[0201] Step 702: Calculate the commodity transaction conversion data to obtain a second estimated value;

[0202] Step 703: performing weighted calculation on the first estimation and the second estimation to obtain a data plane estimation;

[0203] Step 704: clustering the predicted recommended product set using a clustering algorithm based on the product interest behavior data and the product transaction conversion data, and calculating the information surface valuation according to the clustering result;

[0204] Step 705: Calculate the knowledge facet valuation by summarizing knowledge based on the regular preference of the to-be-recommended user for each commodity in the predicted recommended commodity set;

[0205] Step 706: Perform weighted calculation on the data aspect valuation, information aspect valuation and knowledge aspect valuation to obtain the trend intention evaluation value.

[0206] In the embodiment of the present specification, the commodity interest behavior data at least includes the number of clicks on the commodity and the duration of browsing the commodity.

[0207] Calculating the commodity interest behavior data to obtain a first valuation further includes:

[0208] The first estimated value is obtained by performing a weighted calculation on the number of clicks on the product and the browsing time of the product.

[0209] The calculation formula is:

[0210] PEV_1=α*CN+β*ST

[0211] Among them, PEV_1 represents the first valuation, α and β represent weights, CN represents the number of clicks on the product, and ST represents the browsing time of the product.

[0212] The commodity transaction conversion data at least includes commodity transaction volume, secondary transaction volume and negative evaluation quantity.

[0213] like Figure 8 As shown, calculating the commodity transaction conversion data to obtain a second estimate further includes:

[0214] Step 801: Calculate the sum of the commodity transaction volume and the secondary transaction volume to obtain the total commodity transaction volume;

[0215] Step 802: Calculate the ratio of the total transaction volume of the commodity to the number of negative evaluations to obtain the second estimated value.

[0216] The calculation formula is:

[0217]

[0218] Among them, PEV_2 represents the second valuation, TV represents the commodity transaction volume, ATV represents the secondary transaction volume, and NEN represents the number of negative evaluations.

[0219] Specifically, the first estimation and the second estimation are weighted and calculated to obtain the formula for the data plane estimation:

[0220] PEV_DATA=a1*PEV_1+a2*PEV_2

[0221] Among them, PEV_DATA represents the data surface valuation, and a1 and a2 represent weights.

[0222] In the embodiment of the present specification, based on the commodity interest behavior data and commodity transaction conversion data, clustering the predicted recommended commodity set using a clustering algorithm, and calculating the information surface valuation according to the clustering result includes:

[0223] The constructed categories include "interested", "needed", "satisfied", "purchased", "understood", "hated", "redundant", "quality issues", etc. Based on user behavior information and product transaction information, a clustering algorithm is used to classify product intent categories.

[0224] Select k categories from among them, belonging to set K, for example, k = 5;

[0225] Calculate the distance dis from all nodes on the information graph I to k categories:

[0226]

[0227] Among them, I represents all attributes on information node i, such as the number of clicks. K represents all attributes that define category k, and a ratio click rate of more than 3 times indicates "interest".

[0228] Then the information entropy is calculated for the event that the product intent category accurately matches the user intent:

[0229]

[0230] Among them, H represents the entropy value, n represents the total number of nodes in the information graph, and p(x i ) indicates that the information graph node represents the probability of an event occurring.

[0231] Then, the information plane valuation is calculated by the formula: PEV_Information = δ*Hmin, where PEV_Information represents the information plane valuation, δ represents the adjustment coefficient, and Hmin represents the minimum entropy value.

[0232] Then, according to the regular preference of the to-be-recommended user for each product in the predicted recommended product set, the knowledge valuation PEV_Knowledge is calculated through knowledge induction. Exemplarily, it can be calculated through association analysis. This is a common knowledge technology and will not be elaborated in the embodiments of this specification.

[0233] The data aspect valuation, information aspect valuation and knowledge aspect valuation are weighted and calculated to obtain the trend intention evaluation value:

[0234] PEV=b1*PEV DATA+b2*PEV_Information+b3*PEV_Knowledge

[0235] Among them, PEV represents the tendency intention evaluation value, and b1, b2, and b3 represent weights.

[0236] According to one embodiment of this specification, Fig. 9 As shown, the trend intention evaluation value exceeds the first threshold, and the method further includes:

[0237] Step 901: determining whether the trend intention evaluation value exceeds a second threshold;

[0238] Step 902: If yes, then replace the standard micro-expression corresponding to the predicted recommended product set in the initial standard micro-expression-product set mapping library with the second training micro-expression.

[0239] In the embodiment of the present specification, the second threshold can be set based on experience. When the trend intention evaluation value exceeds the second threshold, it means that the recommended product set is highly consistent with the user's needs. At this time, the user's second training micro-expression is more typical. Therefore, the embodiment of the present specification replaces the standard micro-expression in the standard micro-expression library with the second training micro-expression.

[0240] Furthermore, if Fig.10 As shown, if the trend intention evaluation value exceeds the second threshold, the method further includes:

[0241] Step 1001: taking a second training micro-expression corresponding to the trend intention evaluation value exceeding the second threshold as a reference micro-expression;

[0242] Step 1002: performing weighted calculation on multiple reference micro-expressions corresponding to the same standard micro-expression according to the corresponding trend intention evaluation values ​​to obtain an updated micro-expression;

[0243] Step 1003: Replace the standard micro-expression with the updated micro-expression.

[0244] In the embodiments of the present specification, a plurality of second training micro-expressions that meet the requirements of replacing standard micro-expressions may be weightedly calculated according to corresponding trend intention evaluation values ​​to obtain updated micro-expressions, and finally the standard micro-expressions may be replaced with the updated micro-expressions.

[0245] According to one embodiment of this specification, Fig.11 As shown, after determining the recommended product set corresponding to the standard micro-expression according to the pre-trained standard micro-expression-product set mapping library of the user to be recommended, the method further includes:

[0246] Step 1011: Eliminate the products that do not match the user information of the user to be recommended from the recommended product set;

[0247] Recommending the products in the recommended product set to the to-be-recommended user further includes:

[0248] Step 1012: Recommend the remaining products in the recommended product set to the user to be recommended.

[0249] In the embodiment of this specification, the recommended product set is determined based on the user's facial expression, so there may be products in the recommended product set that do not match the user information. Although the user's intention is high, the user cannot order the product. Therefore, the embodiment of this specification removes the products that do not match the user information from the recommended product set, and recommends the remaining products in the recommended product set to the user.

[0250] Exemplarily, the user information of the user may include identity information, asset information, credit information, etc., which is not limited in the embodiments of this specification.

[0251] Based on the same inventive concept, the embodiment of this specification also provides a product recommendation device based on user facial expressions, such as Fig.12 As shown, the device comprises:

[0252] The facial expression information acquisition unit 1201 is used to acquire the facial expression information of the user to be recommended;

[0253] An actual micro-expression calculation unit 1202, configured to calculate the actual micro-expression of the user to be recommended according to the facial expression information;

[0254] A standard micro-expression determination unit 1203 is used to determine a standard micro-expression corresponding to the actual micro-expression according to a pre-trained difference feature model;

[0255] The recommended product set determining unit 1204 is used to determine the recommended product set corresponding to the standard micro-expression according to the pre-trained standard micro-expression-product set mapping library of the user to be recommended;

[0256] The product recommendation unit 1205 is used to recommend the products in the recommended product set to the user to be recommended.

[0257] Furthermore, the product recommendation device based on user facial expression also includes a difference feature model training unit, which is used to:

[0258] Acquire multiple facial expression information of multiple training users and respectively calculate multiple first training micro-expressions corresponding to the multiple facial expression information of each training user and multiple emotion information corresponding to the multiple facial expression information;

[0259] Determine at least one standard micro-expression corresponding to each emotional information of each training user according to a pre-constructed standard micro-expression-emotion library, wherein the standard micro-expression-emotion library includes a correspondence between a plurality of standard micro-expressions and emotional information;

[0260] Calculate the difference between each standard micro-expression and the corresponding first training micro-expression;

[0261] The difference feature model is constructed according to the multiple differences of each training user.

[0262] Furthermore, respectively calculating the difference between each standard micro-expression and the corresponding first training micro-expression further includes:

[0263] Vectorizing the first training micro-expression to obtain a first feature vector;

[0264] Vectorizing the standard micro-expression to obtain a second feature vector;

[0265] The difference feature between the first feature vector and the second feature vector is calculated to obtain the difference between the corresponding standard micro-expression and the first training micro-expression.

[0266] Furthermore, the formula for calculating the difference feature between the first feature vector and the second feature vector is:

[0267]

[0268] Wherein, DF represents the difference feature, DF1 represents the first feature vector, and DF2 represents the second feature vector.

[0269] Furthermore, constructing the difference feature model according to the multiple differences of each training user further includes:

[0270] Calculating a linear relationship between a first training micro-expression and a standard micro-expression of each training user according to a plurality of difference feature vectors of each training user;

[0271] The linear relationship is used as the difference feature model.

[0272] Furthermore, the product recommendation device based on user facial expressions also includes a standard micro-expression-product set mapping library training unit, which is used to:

[0273] Constructing an initial standard micro-expression-product set mapping library for the user to be recommended;

[0274] Acquire facial expression information of the user to be recommended, and calculate a second training micro-expression of the user to be recommended;

[0275] Determine, according to a pre-trained difference feature model, a standard micro-expression corresponding to the second training micro-expression of the user to be recommended;

[0276] Determine a predicted recommended commodity set corresponding to the standard micro-expression of the user to be recommended according to the initial standard micro-expression-commodity set mapping library, and recommend the predicted recommended commodity set to the user to be recommended;

[0277] Acquiring operation information of the to-be-recommended user on the predicted recommended product set, and calculating a trend intention evaluation value according to the operation information;

[0278] Determining whether the trend intention evaluation value exceeds a first threshold;

[0279] If not, adjusting the correspondence between the standard micro-expression and the commodity set in the initial standard micro-expression-commodity set mapping library of the user to be recommended, and repeating the steps of obtaining the facial expression information of the second training user and calculating the second training micro-expression of the second training user;

[0280] If so, a trained standard micro-expression-product set mapping library is obtained.

[0281] Furthermore, the operation information at least includes:

[0282] Product interest behavior data and product transaction conversion data.

[0283] Furthermore, calculating the trend intention evaluation value according to the operation information further includes:

[0284] Calculating the commodity interest behavior data to obtain a first valuation;

[0285] Calculating the commodity transaction conversion data to obtain a second estimated value;

[0286] Performing weighted calculation on the first estimate and the second estimate to obtain a data plane estimate;

[0287] Based on the commodity interest behavior data and commodity transaction conversion data, clustering the predicted recommended commodity set using a clustering algorithm, and calculating the information surface valuation according to the clustering result;

[0288] Calculating the knowledge valuation through knowledge induction according to the regular preference of the to-be-recommended user for each commodity in the predicted recommended commodity set;

[0289] The data aspect valuation, information aspect valuation and knowledge aspect valuation are weightedly calculated to obtain the trend intention evaluation value.

[0290] Furthermore, the product interest behavior data includes at least the number of product clicks and the product browsing time.

[0291] Furthermore, calculating the commodity interest behavior data to obtain a first valuation further includes:

[0292] The first estimated value is obtained by performing a weighted calculation on the number of clicks on the product and the browsing time of the product.

[0293] Furthermore, the commodity transaction conversion data at least includes commodity transaction volume, secondary transaction volume and number of negative reviews.

[0294] Furthermore, calculating the commodity transaction conversion data to obtain a second estimated value further includes:

[0295] Calculating the sum of the commodity transaction volume and the secondary transaction volume to obtain the total commodity transaction volume;

[0296] The ratio of the total commodity transaction volume to the number of negative evaluations is calculated to obtain the second estimated value.

[0297] Furthermore, the product recommendation device based on user facial expressions also includes a standard micro-expression updating unit, which is used to:

[0298] If the trend intention evaluation value exceeds the first threshold, determining whether the trend intention evaluation value exceeds the second threshold;

[0299] If so, the standard micro-expression corresponding to the predicted recommended product set in the initial standard micro-expression-product set mapping library is replaced with the second training micro-expression.

[0300] Furthermore, the standard micro-expression updating unit is further used to use the second training micro-expression corresponding to the trend intention evaluation value exceeding the second threshold as a reference micro-expression;

[0301] Performing weighted calculation on multiple reference micro-expressions corresponding to the same standard micro-expression according to the corresponding trend intention evaluation values ​​to obtain an updated micro-expression;

[0302] The standard micro-expression is replaced with the updated micro-expression.

[0303] Furthermore, the product recommendation device based on user facial expressions also includes a commodity rejection unit, which is used to:

[0304] Eliminate products that do not match the user information of the user to be recommended from the recommended product set;

[0305] Recommending the products in the recommended product set to the to-be-recommended user further includes:

[0306] The remaining commodities in the recommended commodity set are recommended to the user to be recommended.

[0307] Since the principle of solving the problem by the above device is similar to that of the above method, the implementation of the above system can refer to the implementation of the above method, and the repeated parts will not be repeated.

[0308] like Fig.13 The figure is a schematic diagram of the structure of a computer device according to an embodiment of the present specification. The computer device according to the embodiment of the present specification can run the method according to the embodiment of the present specification.

[0309] The computer device 1302 may include one or more processing devices 1304, such as one or more central processing units (CPUs), each of which may implement one or more hardware threads. The computer device 1302 may also include any storage resources 1306 for storing any type of information such as code, settings, data, etc. Without limitation, for example, the storage resources 1306 may include any one or more combinations of the following: any type of RAM, any type of ROM, a flash memory device, a hard disk, an optical disk, etc. More generally, any storage resource may use any technology to store information.

[0310] Further, any storage resource may provide volatile or non-volatile retention of information.

[0311] Further, any storage resource may represent a fixed or removable component of the computer device 1302. In one case, when the processing device 1304 executes the associated instructions stored in any storage resource or combination of storage resources, the computer device 1302 may perform any operation of the associated instructions. The computer device 1302 also includes one or more drive systems 1308 for interacting with any storage resource, such as a hard disk drive system, an optical disk drive system, etc.

[0312] The computer device 1302 may also include an input / output module 1310 (I / O) for receiving various inputs (via input devices 1312) and for providing various outputs (via output devices 1314). A specific output mechanism may include a presentation device 1316 and an associated graphical user interface (GUI) 1318. In other embodiments, the input / output module 1310 (I / O), input device 1312, and output device 1314 may not be included, and the computer device 1302 may be used as a computer device in a network. The computer device 1302 may also include one or more network interfaces 1320 for exchanging data with other devices via one or more communication links 1322. One or more communication buses 1324 couple the components described above together.

[0313] The communication link 1322 may be implemented in any manner, for example, through a local area network, a wide area network (e.g., the Internet), a point-to-point connection, etc., or any combination thereof. The communication link 1322 may include any combination of hardwired links, wireless links, routers, gateway functions, name servers, etc., governed by any protocol or combination of protocols.

[0314] The embodiments of the present specification also provide a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program implements the above method when executed by a processor.

[0315] The embodiments of the present specification also provide a computer-readable instruction, wherein when a processor executes the instruction, the program therein causes the processor to execute the above method.

[0316] It should be understood that in the various embodiments of the present specification, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present specification.

[0317] It should also be understood that in the embodiments of this specification, the term "and / or" is only a description of the association relationship of the associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in the embodiments of this specification generally indicates that the associated objects before and after are in an "or" relationship.

[0318] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the embodiments of this specification can be implemented with electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. 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 exceed the scope of the embodiments of this specification.

[0319] 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.

[0320] In the several embodiments provided in the embodiments of this specification, 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 the units is only a logical function division. 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. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, or it can be an electrical, mechanical or other form of connection.

[0321] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of this specification.

[0322] In addition, each functional unit in each embodiment of the present specification can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of software functional units.

[0323] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of this specification is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the embodiment of this specification. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code.

[0324] The embodiments of this specification use specific embodiments to illustrate the principles and implementation methods of the embodiments of this specification. The description of the above embodiments is only used to help understand the methods and core ideas of the embodiments of this specification. At the same time, for those skilled in the art, according to the ideas of the embodiments of this specification, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on the embodiments of this specification.

Claims

1. A product recommendation method based on user facial expressions, characterized in that: The method comprises: Obtain facial expression information of the user to be recommended; Calculating the actual micro-expression of the user to be recommended according to the facial expression information; Determine a standard micro-expression corresponding to the actual micro-expression according to a pre-trained difference feature model; Determine the recommended product set corresponding to the standard micro-expression according to the pre-trained standard micro-expression-product set mapping library of the user to be recommended; Recommend the products in the recommended product set to the user to be recommended.

2. The method according to claim 1, characterized in that The steps of training the difference feature model include: Acquire multiple facial expression information of multiple training users and respectively calculate multiple first training micro-expressions corresponding to the multiple facial expression information of each training user and multiple emotion information corresponding to the multiple facial expression information; Determine at least one standard micro-expression corresponding to each emotional information of each training user according to a pre-constructed standard micro-expression-emotion library, wherein the standard micro-expression-emotion library includes a correspondence between a plurality of standard micro-expressions and emotional information; Calculate the difference between each standard micro-expression and the corresponding first training micro-expression; The difference feature model is constructed according to the multiple differences of each training user.

3. The method according to claim 2, characterized in that Calculating the difference between each standard micro-expression and the corresponding first training micro-expression further includes: Vectorizing the first training micro-expression to obtain a first feature vector; Vectorizing the standard micro-expression to obtain a second feature vector; The difference feature between the first feature vector and the second feature vector is calculated to obtain the difference between the corresponding standard micro-expression and the first training micro-expression.

4. The method according to claim 3, characterized in that The formula for calculating the difference feature between the first eigenvector and the second eigenvector is: Wherein, DF represents the difference feature, DF1 represents the first feature vector, and DF2 represents the second feature vector.

5. The method according to claim 3, characterized in that: Constructing the difference feature model according to the multiple differences of each training user further includes: Calculating a linear relationship between a first training micro-expression and a standard micro-expression of each training user according to a plurality of difference feature vectors of each training user; The linear relationship is used as the difference feature model.

6. The method according to claim 1, characterized in that Training the to-be-recommended user standard micro-expression-product set mapping library further includes: Constructing an initial standard micro-expression-product set mapping library for the user to be recommended; Acquire facial expression information of the user to be recommended, and calculate a second training micro-expression of the user to be recommended; Determine, according to a pre-trained difference feature model, a standard micro-expression corresponding to the second training micro-expression of the user to be recommended; Determine a predicted recommended commodity set corresponding to the standard micro-expression of the user to be recommended according to the initial standard micro-expression-commodity set mapping library, and recommend the predicted recommended commodity set to the user to be recommended; Acquiring operation information of the to-be-recommended user on the predicted recommended product set, and calculating a trend intention evaluation value according to the operation information; Determining whether the trend intention evaluation value exceeds a first threshold; If not, adjusting the correspondence between the standard micro-expression and the commodity set in the initial standard micro-expression-commodity set mapping library of the user to be recommended, and repeating the steps of obtaining the facial expression information of the second training user and calculating the second training micro-expression of the second training user; If so, a trained standard micro-expression-product set mapping library is obtained.

7. The method according to claim 6, characterized in that The operation information at least includes: Product interest behavior data and product transaction conversion data.

8. The method according to claim 7, characterized in that Calculating the trend intention evaluation value according to the operation information further includes: Calculating the commodity interest behavior data to obtain a first valuation; Calculating the commodity transaction conversion data to obtain a second estimated value; Performing weighted calculation on the first estimate and the second estimate to obtain a data plane estimate; Based on the commodity interest behavior data and commodity transaction conversion data, clustering the predicted recommended commodity set using a clustering algorithm, and calculating the information surface valuation according to the clustering result; Calculating the knowledge valuation through knowledge induction according to the regular preference of the to-be-recommended user for each commodity in the predicted recommended commodity set; The data aspect valuation, information aspect valuation and knowledge aspect valuation are weightedly calculated to obtain the trend intention evaluation value.

9. The method according to claim 8, characterized in that The product interest behavior data at least includes the number of product clicks and the product browsing time.

10. The method according to claim 9, characterized in that Calculating the commodity interest behavior data to obtain a first valuation further includes: The first estimated value is obtained by performing a weighted calculation on the number of clicks on the product and the browsing time of the product.

11. The method according to claim 8, characterized in that The commodity transaction conversion data at least includes commodity transaction volume, secondary transaction volume and negative evaluation quantity.

12. The method according to claim 11, characterized in that Calculating the commodity transaction conversion data to obtain a second estimated value further includes: Calculating the sum of the commodity transaction volume and the secondary transaction volume to obtain the total commodity transaction volume; The ratio of the total commodity transaction volume to the number of negative evaluations is calculated to obtain the second estimated value.

13. The method according to claim 6, characterized in that If the trend intention evaluation value exceeds the first threshold, the method further includes: Determining whether the trend intention evaluation value exceeds a second threshold; If so, the standard micro-expression corresponding to the predicted recommended product set in the initial standard micro-expression-product set mapping library is replaced with the second training micro-expression.

14. The method according to claim 13, characterized in that If the trend intention evaluation value exceeds the second threshold, the method further includes: taking a second training micro-expression corresponding to the trend intention evaluation value exceeding the second threshold as a reference micro-expression; Performing weighted calculation on multiple reference micro-expressions corresponding to the same standard micro-expression according to the corresponding trend intention evaluation values ​​to obtain an updated micro-expression; The standard micro-expression is replaced with the updated micro-expression.

15. The method according to claim 1, characterized in that After determining the recommended product set corresponding to the standard micro-expression according to the pre-trained standard micro-expression-product set mapping library of the user to be recommended, the method further includes: Eliminate products that do not match the user information of the user to be recommended from the recommended product set; Recommending the products in the recommended product set to the to-be-recommended user further includes: The remaining commodities in the recommended commodity set are recommended to the user to be recommended.

16. A product recommendation device based on user facial expressions, characterized in that: The device comprises: A facial expression information acquisition unit, used to acquire facial expression information of the user to be recommended; An actual micro-expression calculation unit, used to calculate the actual micro-expression of the user to be recommended according to the facial expression information; A standard micro-expression determination unit, used to determine a standard micro-expression corresponding to the actual micro-expression according to a pre-trained difference feature model; A recommended product set determination unit, configured to determine a recommended product set corresponding to the standard micro-expression according to a pre-trained standard micro-expression-product set mapping library of the user to be recommended; The commodity recommendation unit is used to recommend the commodities in the recommended commodity set to the user to be recommended.

17. A computer device comprising a memory, a processor, and a computer program stored in the memory, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 15 is implemented.

18. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 15 is implemented.

19. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 15 is implemented.

Images