Product recommendation method and device, electronic equipment and storage medium

Abstract

The invention provides a product recommendation method and device, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence. The method comprises the following steps: determining historical browsing information of a first object for each first candidate product, product preference information of the first object for each second candidate product, and description feedback information of the first object for each third candidate product; selecting a reference candidate product from the plurality of first candidate products according to the historical browsing information, selecting a reference candidate product from the plurality of second candidate products according to the product preference information, and selecting a reference candidate product from the plurality of third candidate products according to the description feedback information; determining a target object group to which the first object belongs according to the network use information and the attribute information of the first object, and selecting at least one target product according to the group preference feature of the target object group and the candidate product feature of the reference candidate product, the technical problem that proper products cannot be accurately recommended to consumers in the prior art is solved.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, and in particular to a product recommendation method, device, electronic device and storage medium. Background Art

[0002] When consumers are faced with a large number of products on shopping platforms, they often compare prices and consider the cost-effectiveness before making a purchase. Due to the differences in purchasing behavior among different categories of consumers, it is necessary to provide users with personalized recommendation services to increase the added value of consumer purchases. In the related art, it is not possible to accurately recommend suitable products to consumers. Summary of the invention

[0003] The present disclosure aims to solve one of the technical problems in the related art at least to some extent.

[0004] To this end, the first objective of the present invention is to propose a product recommendation method to improve the accuracy of product recommendations.

[0005] The second objective of the present disclosure is to provide a product recommendation device.

[0006] A third objective of the present disclosure is to provide an electronic device.

[0007] A fourth objective of the present disclosure is to provide a computer-readable storage medium.

[0008] A fifth object of the present disclosure is to provide a computer program product.

[0009] To achieve the above-mentioned purpose, the first aspect of the present disclosure proposes a product recommendation method, including: determining the historical browsing information of the first object for each first candidate product, determining the product preference information of the first object for each second candidate product, and determining the description feedback information of the first object for each third candidate product; selecting a reference candidate product from multiple first candidate products according to the historical browsing information, selecting a reference candidate product from multiple second candidate products according to the product preference information, and selecting a reference candidate product from multiple third candidate products according to the description feedback information, wherein the reference candidate product has corresponding candidate product characteristics; determining the network usage information and attribute information of the first object; determining the target object group to which the first object belongs according to the network usage information and the attribute information, wherein the target object group has a corresponding group preference characteristic, and the group preference characteristic is used to describe the preference of the target object group for the product; and selecting at least one target product from some of the reference candidate products according to the group preference characteristic and the candidate product characteristic.

[0010] To achieve the above-mentioned purpose, the second aspect of the present disclosure proposes a product recommendation device, including: a first determination module, used to determine the historical browsing information of the first object for each first candidate product, determine the product preference information of the first object for each second candidate product, and determine the description feedback information of the first object for each third candidate product; a selection module, used to select a reference candidate product from multiple first candidate products according to the historical browsing information, select a reference candidate product from multiple second candidate products according to the product preference information, and select a reference candidate product from multiple third candidate products according to the description feedback information, wherein the reference candidate product has corresponding candidate product features; a second determination module, used to determine the network usage information and attribute information of the first object; a third determination module, used to determine the target object group to which the first object belongs according to the network usage information and the attribute information, wherein the target object group has a corresponding group preference feature, and the group preference feature is used to describe the preference of the target object group for the product; and a recommendation module, used to select at least one target product from some of the reference candidate products according to the group preference feature and the candidate product feature.

[0011] To achieve the above-mentioned purpose, the third aspect embodiment of the present disclosure proposes an electronic device, including: a processor, and a memory communicatively connected to the processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory to implement a product recommendation method proposed in the first aspect embodiment of the present disclosure.

[0012] To achieve the above-mentioned purpose, the fourth aspect embodiment of the present disclosure proposes a computer-readable storage medium, in which computer execution instructions are stored. When the computer execution instructions are executed by a processor, they are used to implement a product recommendation method proposed in the first aspect embodiment of the present disclosure.

[0013] To achieve the above-mentioned purpose, the fifth aspect of the present disclosure proposes a computer program product, including a computer program, which, when executed by a processor, implements the product recommendation method proposed in the first aspect of the present disclosure.

[0014] The product recommendation method, device, electronic device, non-transitory computer-readable storage medium storing computer instructions, and computer program product provided by the present disclosure determine the historical browsing information of the first object for each first candidate product, determine the product preference information of the first object for each second candidate product, and determine the description feedback information of the first object for each third candidate product, and select reference candidate products from multiple first candidate products according to the historical browsing information, select reference candidate products from multiple second candidate products according to the product preference information, and select reference candidate products from multiple third candidate products according to the description feedback information, wherein the reference candidate products have corresponding candidate product features, determine the network usage information and attribute information of the first object, determine the target object group to which the first object belongs according to the network usage information and the attribute information, wherein the target object group has a corresponding group preference feature, and the group preference feature is used to describe the preference of the target object group for the product, and select at least one target product from some of the reference candidate products according to the group preference feature and the candidate product feature. Therefore, not only can product screening be performed based on the personalized situation of the first object, but product screening can also be further performed based on the characteristics of the target object group to which the first object belongs, so that the target products obtained by screening can be more in line with the preferences of the first object, thereby improving the accuracy of product recommendations.

[0015] Additional aspects and advantages of the present disclosure will be given in part in the following description and in part will be obvious from the following description or learned through practice of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The above and / or additional aspects and advantages of the present disclosure will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0017] Figure 1 A flowchart of a product recommendation method provided by an embodiment of the present disclosure;

[0018] Figure 2 A flowchart of another product recommendation method provided by an embodiment of the present disclosure;

[0019] Figure 3 is a structural diagram of a personalized shopping recommendation system based on machine learning in an embodiment of the present disclosure;

[0020] Figure 4 A schematic diagram of the structure of a product recommendation device provided by an embodiment of the present disclosure;

[0021] Figure 5 It is a schematic structural diagram of an electronic device shown in an exemplary embodiment of the present disclosure. DETAILED DESCRIPTION

[0022] Embodiments of the present disclosure are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present disclosure, and should not be construed as limiting the present disclosure.

[0023] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.) and signals involved in this disclosure are all authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions. The acquisition, transmission, storage, use and processing of data in the technical solution of this disclosure are in compliance with the relevant provisions of national laws and regulations.

[0024] It should be noted that in the embodiments of the present disclosure, certain software, components, models and other existing solutions in the industry may be mentioned, which should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present disclosure, but it does not mean that the applicant has or will necessarily use the solution.

[0025] Figure 1 A flowchart of a product recommendation method provided in an embodiment of the present disclosure.

[0026] This embodiment takes the product recommendation method being configured in a product recommendation device as an example. In this embodiment, the product recommendation method can be configured in the product recommendation device, and the product recommendation device can be set in a server or in an electronic device, without limitation.

[0027] This embodiment takes the product recommendation method configured in an electronic device as an example, wherein the electronic device is a hardware device with various operating systems, such as a smart phone, a tablet computer, a personal digital assistant, an e-book, etc.

[0028] It should be noted that the execution subject of the embodiment of the present disclosure may be, for example, a central processing unit (CPU) in a server or electronic device in terms of hardware, or may be, for example, a related background service in a server or electronic device in terms of software, without limitation.

[0029] like Figure 1 As shown, the recommended method for this product includes:

[0030] S101: Determine the first subject's historical browsing information for each first candidate product, determine the first subject's product preference information for each second candidate product, and determine the first subject's description feedback information for each third candidate product.

[0031] The first object may refer to a consumer to whom a product is to be recommended. The number of the first objects may be one or more.

[0032] In some embodiments, the historical browsing information of the first subject for each first candidate product can be determined, wherein the first candidate product can be a candidate product to be screened based on the dimension of the historical browsing information, and the historical browsing information is used to describe the historical browsing of the first candidate product by the first subject, such as browsing frequency, browsing duration, etc. The number of first candidate products can be one or more. The browsing data of the first subject in the shopping platform can be collected, and one or more first candidate products that the first subject focused on browsing can be determined based on the browsing data, and the historical browsing information of the first subject for each first candidate product can be obtained by parsing the browsing log in the background, without limitation.

[0033] In some embodiments, the historical browsing information may be used as reference information when screening reference candidate products from among a plurality of first candidate products.

[0034] In some embodiments, the product preference information of the first subject for each second candidate product can be determined, wherein the second candidate product can be a candidate product to be screened based on the dimension of the product preference information, and the product preference information is used to describe the first subject's preference for the second candidate product, such as whether to purchase, the quantity purchased, the number of purchases, etc. The number of second candidate products can be one or more. The actual purchase data of the first subject in the shopping platform can be collected, and one or more second candidate products that the first subject prefers to purchase can be determined based on the actual purchase data, and the product preference information of the first subject for each second candidate product can be obtained by parsing the purchase log in the background, without limitation.

[0035] In some embodiments, the product preference information may be used as reference information when screening reference candidate products from a plurality of second candidate products.

[0036] In some embodiments, the description feedback information of the first subject on each third candidate product can be determined, wherein the third candidate product can be a candidate product to be screened based on the dimension of the description feedback information, and the description feedback information is used to describe the evaluation feedback of the first subject on the third candidate product, such as quality feedback on the third candidate product, description feedback on the merchant to which the third candidate product belongs, etc. The number of third candidate products can be one or more. The actual evaluation data of the first subject in the shopping platform can be collected, and one or more third candidate products that the first subject has evaluated and fed back can be determined based on the actual evaluation data, and the evaluation content, feedback information, etc. of the first subject on each third candidate product can be parsed from the background system log as the description feedback information, and there is no limitation on this.

[0037] In some embodiments, the description feedback information may be used as reference information when screening reference candidate products from a plurality of third candidate products.

[0038] Among them, the first candidate product, the second candidate product, and the third candidate product may be the same candidate product, or partially the same candidate product, or different candidate products, and there is no restriction on this. In other words, multiple first candidate products and multiple second candidate products may have overlapping candidate products, or completely different candidate products, or completely the same candidate products, multiple first candidate products and multiple third candidate products may have overlapping candidate products, or completely different candidate products, or completely the same candidate products, and multiple second candidate products and multiple third candidate products may have overlapping candidate products, or completely different candidate products, or completely the same candidate products, and there is no restriction on this.

[0039] S102: Selecting a reference candidate product from a plurality of first candidate products according to historical browsing information, selecting a reference candidate product from a plurality of second candidate products according to product preference information, and selecting a reference candidate product from a plurality of third candidate products according to description feedback information, wherein the reference candidate product has corresponding candidate product features.

[0040] In the above, when determining the historical browsing information of the first object for each first candidate product, a reference candidate product can be selected from multiple first candidate products according to the historical browsing information. For example, the historical browsing information can be analyzed with reference to some screening rules to determine the first candidate products corresponding to the historical browsing information that meets the screening rules as the reference candidate products, or the historical browsing information can be statistically analyzed to determine the first candidate products that meet some statistical rules as the reference candidate products, and there is no limitation on this.

[0041] After determining the product preference information of the first object for each second candidate product, the product preference information can be subjected to feature analysis to determine the product preference features of the first object for each second candidate product, and the second candidate products whose product preference features meet the preset features can be screened as reference candidate products. Alternatively, the first object's preference for each second candidate product can be scored based on the product preference information to obtain a preference scoring result, and suitable second candidate products can be screened as reference candidate products based on the preference scoring result, and there is no restriction on this.

[0042] After determining the description feedback information of the first object on each third candidate product, semantic recognition can be performed on the description feedback information to determine the semantics and context information of the description feedback, and the third candidate products can be screened as reference candidate products based on the semantics and context information of the description feedback, without limitation.

[0043] In some embodiments, the number of reference candidate products screened from multiple first candidate products may be at least one, the number of reference candidate products screened from multiple second candidate products may be at least one, the number of reference candidate products screened from multiple third candidate products may be at least one, and the reference candidate products screened based on different dimensions (the dimension of historical browsing information, or the dimension of product preference information, or the dimension of description feedback information) may be the same or different. Therefore, the number of reference candidate products screened from multiple first candidate products, multiple second candidate products, and multiple third candidate products may be multiple, and there is no limitation on this.

[0044] Therefore, by selecting reference candidate products from multiple first candidate products based on historical browsing information, selecting reference candidate products from multiple second candidate products based on product preference information, and selecting reference candidate products from multiple third candidate products based on description feedback information, wherein the reference candidate products have corresponding candidate product features, a number of reference candidate products are obtained by screening based on multiple dimensions. This screening process can be understood as a preliminary screening process. In the preliminary screening process, the comprehensiveness of the reference candidate product selection is effectively improved, thereby supporting the improvement of the accuracy of subsequent product recommendations.

[0045] S103: Determine network usage information and attribute information of the first object.

[0046] After screening and obtaining multiple reference candidate products, the network usage information and attribute information of the first object can be determined, wherein the network usage information and attribute information can include network usage information and attribute information, the network usage information is used to describe the network usage of the first object, the attribute information is used to describe the characteristic attributes of the first object, the network usage information and attribute information can be used to determine the grouping to which the first object belongs, such as determining which grouping the first object belongs to, and the grouping to which the first object belongs can be called the target object group.

[0047] For example, the network usage information may be, for example, the network access time, access duration, access content information, etc. of the first object, and the attribute information may be, for example, the attribute characteristics of the first object, which is not limited.

[0048] In some embodiments, in the process of determining the network usage information of the first object, the network access address of the first object may be obtained, and access log data matching the network access address may be determined in the background, and the access log data may be analyzed to determine the network usage information. Alternatively, the account information of the first object may be monitored, and the access content and access features of the account information on the network may be counted, and the network usage information may be determined based on the access content and access features, and there is no limitation on this.

[0049] In some embodiments, the attribute information may be pre-registered by the first object and stored in the electronic device, then the first object identifier may be obtained based on the first object, and the candidate attribute information corresponding to the first object identifier may be used as the attribute information of the first object. Alternatively, the attribute information input by the first object may be received in real time, or the attribute information of the first object may be determined based on any other possible method, without limitation.

[0050] S104: Determine the target object group to which the first object belongs according to the network usage information and the attribute information, wherein the target object group has a corresponding group preference feature, and the group preference feature is used to describe the preference of the target object group for the product.

[0051] After determining the network usage information and attribute information of the first object, the target object group to which the first object belongs can be determined based on the network usage information and attribute information. The target object group can be one of multiple candidate object groups. Multiple candidate object groups can be pre-classified, and different candidate object groups have different network usage information and attribute information. The group preference characteristics of each candidate object group can be pre-marked, and the group preference characteristics of the candidate object group that are the same as the target object group can be called the group preference characteristics corresponding to the target object group. The group preference characteristics can be used to describe the target object group's preference for products.

[0052] That is to say, in the embodiments of the present disclosure, the target object group to which the first object belongs can be first determined based on the network usage information and attribute information of the first object, and group preference characteristics can be pre-labeled for the target object group, and the group preference characteristics are used to describe the target object group's preferences for products. Then, the identified group preference characteristics of the target object group can be used to further screen the several reference candidate products obtained in the above initial screening. Thus, not only can product screening be performed based on the personalized situation of the first object, but product screening can also be further performed based on the characteristics of the target object group to which the first object belongs, so that the target products obtained by screening can be more in line with the preferences of the first object, and the accuracy of product recommendations can be improved.

[0053] S105: Select at least one target product from some reference candidate products according to the group preference characteristics and the candidate product characteristics.

[0054] After determining the candidate product features of each of the plurality of reference candidate products and determining the group preference features of the target object group to which the first object belongs, the group preference features and the candidate product features can be used to select at least one target product from some of the reference candidate products. The selected target product can be a preferred product for recommendation to the first object.

[0055] In some embodiments, a recommendation score can be calculated for each reference candidate product based on the candidate product features and the group preference features, and one or more target products can be selected from several reference candidate products based on the recommendation score; or a selection model (such as a neural network model or a machine learning model) can be trained and the selection model can be used to process the candidate product features and the group preference features to select at least one target product from several reference candidate products; or any other possible method can be used to select at least one target product from some reference candidate products based on the group preference features and the candidate product features, without limitation.

[0056] In this embodiment, by determining the historical browsing information of the first object for each first candidate product, determining the product preference information of the first object for each second candidate product, and determining the description feedback information of the first object for each third candidate product, and selecting a reference candidate product from multiple first candidate products according to the historical browsing information, selecting a reference candidate product from multiple second candidate products according to the product preference information, and selecting a reference candidate product from multiple third candidate products according to the description feedback information, wherein the reference candidate product has corresponding candidate product characteristics, determining the network usage information and attribute information of the first object, determining the target object group to which the first object belongs according to the network usage information and attribute information, wherein the target object group has a corresponding group preference feature, the group preference feature is used to describe the preference of the target object group for the product, and selecting at least one target product from some reference candidate products according to the group preference feature and the candidate product feature. Thus, not only can product screening be achieved based on the personalized situation of the first object, but also product screening can be further performed based on the characteristic situation of the target object group to which the first object belongs, so that the screened target product can be more in line with the preference of the first object, and the accuracy of product recommendation can be improved.

[0057] In some embodiments of the present disclosure, in the process of determining the historical browsing information of each first candidate product by the first object, the initial browsing information of each first candidate product can be determined, and the first browsing information with abnormalities is identified from multiple initial browsing information, and the browsing characteristics of the remaining initial browsing information are analyzed to obtain browsing characteristic information, and the first initial browsing information is corrected according to the browsing characteristic information to obtain the second browsing information, and the remaining initial browsing information and the second browsing information are deduplicated to obtain historical browsing information. Thus, the initial browsing information initially collected is subjected to abnormality detection and deduplication processing, and the second browsing information obtained by abnormality detection and processing and the initial browsing information without abnormalities are deduplicated, and the browsing information obtained by deduplication processing is used as historical browsing information, which can effectively ensure the data quality of the historical browsing information, so that when the reference candidate product is screened from multiple first candidate products based on the historical browsing information, the selection accuracy of the reference candidate product can be improved, and the initial screening effect of the reference candidate product can be improved.

[0058] In some embodiments of the present disclosure, in the process of determining the product preference information of each second candidate product of the first object, it can be to determine the reference product characteristics of each second candidate product, determine the initial preference information of the first object for each second candidate product, and identify the first preference information with abnormalities from multiple initial preference information, wherein the first preference information has a corresponding second candidate product, and the first preference information is corrected according to the reference product characteristics of the second candidate product corresponding to the remaining initial preference information and the reference product characteristics of the second candidate product corresponding to the first preference information to obtain the second preference information, and the remaining initial preference information and the second preference information are deduplicated to obtain the product preference information. Thus, the initial preference information initially collected is subjected to abnormality detection and deduplication, and the second preference information obtained by abnormality detection and processing and the initial preference information without abnormality are deduplicated, and the preference information obtained by deduplication is used as the product preference information, which can effectively ensure the data quality of the product preference information, so that when the reference candidate product is screened from multiple second candidate products based on the product preference information, the selection accuracy of the reference candidate product can be improved, and the initial screening effect of the reference candidate product can be improved.

[0059] In some embodiments of the present disclosure, in the process of selecting a reference candidate product from multiple first candidate products based on historical browsing information, the browsing duration, browsing time, and browsing frequency of each first candidate product by the first object may be determined based on the historical browsing information, and browsing score information may be determined based on the browsing duration, browsing time, and browsing frequency, wherein the browsing score information is positively correlated with the browsing stickiness of the first object to the first candidate product, and the reference candidate product may be selected from multiple first candidate products based on the browsing score information. In this way, the reference candidate product may be accurately selected from multiple first candidate products.

[0060] For example, the browsing characteristics of the first object for each first candidate product can be analyzed based on historical browsing information, such as browsing duration, browsing time and browsing frequency, and then the browsing behavior of the first object for the first candidate product can be scored based on the browsing characteristics to obtain browsing score information. The higher the browsing score information, the more the first object tends to like browsing (higher browsing stickiness) the first candidate product, and the lower the browsing score information, the less the first object likes browsing (lower browsing stickiness) the first candidate product. After calculating the browsing score information of the first object for each first candidate product, the browsing score information with a higher score value can be selected from multiple browsing score information, and the first candidate product to which the browsing score information with a higher score value belongs can be used as a reference candidate product.

[0061] In some embodiments of the present disclosure, in the process of selecting a reference candidate product from multiple second candidate products according to product preference information, the purchase frequency, purchase quantity and purchase time of each second candidate product by the first subject may be determined according to the product preference information, and preference score information may be determined according to the purchase frequency, purchase quantity and purchase time, wherein the preference score information is positively correlated with the first subject's preference for the second candidate product, and the reference candidate product may be selected from multiple second candidate products according to the preference score information. Thus, the reference candidate product may be accurately selected from multiple second candidate products.

[0062] For example, the preference characteristics of the first subject for each second candidate product can be analyzed based on the product preference information, such as the purchase frequency, purchase quantity and purchase time, and then the preference of the first subject for the second candidate product can be scored based on the preference characteristics to obtain preference score information. The higher the preference score information, the more the first subject tends to prefer to buy (higher preference) the second candidate product, and the lower the preference score information, the less the first subject likes to buy (lower preference) the second candidate product. After calculating the preference score information of the first subject for each second candidate product, preference score information with a higher score value can be selected from multiple preference score information, and the second candidate product to which the preference score information with a higher score value belongs can be used as a reference candidate product.

[0063] In some embodiments of the present disclosure, in the process of selecting a reference candidate product from multiple third candidate products according to description feedback information, semantic recognition may be performed on each description feedback information to obtain description semantic features, and description tendency information corresponding to each description semantic feature is determined, wherein the description tendency information includes: positive feedback description or negative feedback description, description scoring information is determined according to the positive feedback description or the negative feedback description, wherein the description scoring information is positively correlated with the degree of positive feedback description, and the reference candidate product is selected from multiple third candidate products according to the description scoring information. Thus, accurate selection of the reference candidate product from multiple third candidate products is achieved.

[0064] For example, the description tendency information of the first object for each third candidate product can be analyzed based on the description feedback information, and the description tendency information is, for example, a positive feedback description or a negative feedback description. Then, the description of the third candidate product by the first object is scored based on the description tendency information to obtain description scoring information. Among them, the higher the description scoring information, the more the first object tends to describe the third candidate product positively (the degree of positive feedback description is higher), and the lower the description scoring information, the more the first object prefers the negative description (the degree of positive feedback description is lower) of the third candidate product. After calculating the description scoring information of each third candidate product by the first object, the description scoring information with a higher scoring value can be selected from multiple description scoring information, and the third candidate product to which the description scoring information with a higher scoring value belongs is used as a reference candidate product.

[0065] Figure 2 A flowchart of another product recommendation method provided in an embodiment of the present disclosure.

[0066] like Figure 2 As shown, the recommended method for this product includes:

[0067] S201: Determine the first subject's historical browsing information for each first candidate product, determine the first subject's product preference information for each second candidate product, and determine the first subject's description feedback information for each third candidate product.

[0068] S202: Select a reference candidate product from a plurality of first candidate products according to historical browsing information, select a reference candidate product from a plurality of second candidate products according to product preference information, and select a reference candidate product from a plurality of third candidate products according to description feedback information, wherein the reference candidate product has corresponding candidate product features.

[0069] S203: Determine network usage information and attribute information of the first object.

[0070] For the description of S201 - S203 , please refer to the above embodiment, which will not be repeated here.

[0071] S204: Determine the target object group to which the first object belongs according to the network usage information and the attribute information, wherein the target object group has a corresponding group preference feature, and the group preference feature is used to describe the preference of the target object group for the product.

[0072] In some embodiments, in the process of determining the target object group to which the first object belongs based on network usage information and attribute information, candidate information of at least one candidate object group may be determined, wherein the candidate information includes: network usage information and attribute information of each candidate object in the candidate object group, and the similarity between the network usage information and attribute information and the candidate information is determined, and the maximum similarity is selected from at least one similarity, and the candidate object group to which the candidate information corresponding to the maximum similarity belongs is used as the target object group.

[0073] After determining the network usage information and attribute information of the first object, the target object group to which the first object belongs can be determined based on the network usage information and attribute information. The target object group can be one of multiple candidate object groups. Multiple candidate object groups can be pre-classified, and different candidate object groups have different network usage information and attribute information. The group preference characteristics of each candidate object group can be pre-marked, and the group preference characteristics of the candidate object group that are the same as the target object group can be called the group preference characteristics corresponding to the target object group. The group preference characteristics can be used to describe the target object group's preference for products.

[0074] After determining the network usage information and attribute information of each candidate object in the candidate object group, the similarity between the network usage information and attribute information of the first object and the network usage information and attribute information of each candidate object can be calculated, and then the maximum similarity is selected from at least one similarity, and the candidate object group to which the candidate information corresponding to the maximum similarity belongs is used as the target object group. Accurate grouping of the first object is achieved, and the recognition accuracy of the group preference feature of the first object is improved, thereby supporting the rationality and accuracy of the target product selection.

[0075] S205: Input the group preference features and the candidate product features into the recommendation model to obtain the recommendation score corresponding to each reference candidate product output by the recommendation model, wherein the recommendation model has learned the mapping relationship between the group preference features, the candidate product features, and the recommendation score corresponding to each reference candidate product.

[0076] That is, the artificial intelligence model can be used to determine the recommendation score of each reference candidate product. The higher the recommendation score, the more suitable the corresponding reference candidate product is for being recommended to the first object, and the lower the recommendation score, the less suitable the corresponding reference candidate product is for being recommended to the first object. The recommendation model can be pre-trained, and the recommendation model can be a neural network model or a machine learning model in artificial intelligence. The recommendation model has learned the mapping relationship between the group preference characteristics, the candidate product characteristics, and the recommendation score corresponding to each reference candidate product.

[0077] In this embodiment, the candidate product features and group preference features of each reference candidate product may be input into the recommendation model, and the recommendation model may score the candidate product features and group preference features of each reference candidate product to obtain a recommendation score corresponding to the reference candidate product.

[0078] S206: Select at least one target product from some reference candidate products according to the recommendation score.

[0079] After obtaining the recommendation score corresponding to each reference candidate product, at least one target product can be selected from some of the reference candidate products according to the recommendation score. For example, multiple recommendation scores can be sorted, and several reference candidate products corresponding to the recommended scores ranked first can be used as target products, or several reference candidate products corresponding to several recommended scores greater than the score threshold can be used as target products, and there is no restriction on this.

[0080] S207: Acquire a recommendation description template, wherein the recommendation description template includes: candidate recommendation information corresponding to the candidate preference features and the reference product features.

[0081] Among them, some candidate preference features and reference product features can be pre-configured in the recommendation description template, and reasonable candidate recommendation information can be set for the configured candidate preference features and reference product features.

[0082] S208: Determine candidate recommendation information corresponding to the group preference characteristics and the candidate product characteristics from the recommendation description template, and use the candidate recommendation information corresponding to the group preference characteristics and the candidate product characteristics as target recommendation information of the target product.

[0083] S209: Provide target recommendation information.

[0084] After selecting at least one target product from some of the reference candidate products, target recommendation information of the recommended target product can also be provided to the first object, thereby enabling the reason why the target product is recommended to be clearly known, thereby supporting the improvement of the probability of the target product being selected by the first object, and further improving the accuracy and recommendation effect of the target product recommendation.

[0085] In this embodiment, by determining the historical browsing information of the first object for each first candidate product, determining the product preference information of the first object for each second candidate product, and determining the description feedback information of the first object for each third candidate product, and selecting a reference candidate product from multiple first candidate products according to the historical browsing information, selecting a reference candidate product from multiple second candidate products according to the product preference information, and selecting a reference candidate product from multiple third candidate products according to the description feedback information, wherein the reference candidate product has corresponding candidate product characteristics, determining the network usage information and attribute information of the first object, determining the target object group to which the first object belongs according to the network usage information and attribute information, wherein the target object group has a corresponding group preference feature, the group preference feature is used to describe the preference of the target object group for the product, and selecting at least one target product from some reference candidate products according to the group preference feature and the candidate product feature. Thus, not only can product screening be achieved based on the personalized situation of the first object, but also product screening can be further performed based on the characteristic situation of the target object group to which the first object belongs, so that the screened target product can be more in line with the preference of the first object, and the accuracy of product recommendation can be improved. After selecting at least one target product from some reference candidate products, target recommendation information for recommending the target product can also be provided to the first object, thereby enabling the first object to clearly understand the reason why the target product is recommended, thereby supporting the increase in the probability of the target product being selected by the first object and further improving the accuracy and recommendation effect of the target product recommendation.

[0086] The above embodiment is described as follows (taking the first object as a user as an example):

[0087] The product recommendation method provided in the embodiments of the present disclosure can be applied in a personalized shopping recommendation system based on machine learning, such as Figure 3 As shown, Figure 33 is a schematic diagram of the structure of a personalized shopping recommendation system based on machine learning in an embodiment of the present disclosure, which may include: a data collection module 31, a data preprocessing module 32, a feature engineering module 33, a model training module 34, a recommendation generation module 35 and a recommendation evaluation module 36. Among them, the data collection module is used to collect data from user interactions, and the data includes but is not limited to the user's browsing history, purchase record, evaluation feedback, and data obtained through social media and website behavior tracking. The data preprocessing module is connected to the data collection module and is used to clean, remove duplicates, process missing values, and format the collected user data to ensure the quality and consistency of the data for subsequent processing. The feature engineering module is connected to the data preprocessing module and is used to extract features related to user interests and behaviors from the preprocessed data, including but not limited to time series analysis, shopping frequency statistics, product category preferences, etc. The model training module is connected to the feature engineering module and uses a machine learning algorithm to train a personalized recommendation model on the extracted feature data set. The recommendation model can be trained based on collaborative filtering, matrix decomposition, deep learning or other suitable algorithms. The recommendation generation module is connected to the model training module. It generates a recommendation list for a specific user based on the trained personalized recommendation model and real-time user feature data. The products in the recommendation list (an optional example of the target product) are sorted according to the predicted user preferences and possible purchase intentions. The recommendation evaluation module is connected to the recommendation generation module and is used to evaluate the effect of the generated recommendation list. The performance of the recommendation system is evaluated by calculating indicators such as the accuracy, recall, and coverage of the recommendation.

[0088] In some embodiments, the data collection module includes a web crawler submodule for periodically crawling product information and user review data from web pages and online marketplaces.

[0089] In some embodiments, the data preprocessing module includes an outlier detection and processing mechanism for identifying and processing abnormal data points that do not conform to the expected distribution.

[0090] In some embodiments, the feature engineering module includes a user grouping function that divides users into different groups based on their historical behaviors and attributes to facilitate more refined feature extraction.

[0091] In some embodiments, the model training module includes a model optimization unit for adjusting model parameters and applying cross-validation and other techniques to improve the generalization ability and accuracy of the model.

[0092] In some embodiments, the recommendation generation module includes an explanation submodule that provides explanations for the reasons for the recommendation to help users understand the logic behind the recommendation.

[0093] In some embodiments, the recommendation evaluation module includes an A / B testing function to compare the effects of different recommendation strategies, thereby optimizing the recommendation algorithm.

[0094] As a result, it is possible to accurately capture and analyze the user's shopping history and behavior patterns, significantly improving the personalization and accuracy of recommendations. The data preprocessing module improves the efficiency and quality of data processing. The feature engineering module deeply explores user interests, the model training module uses multiple algorithms to optimize the recommendation model, the recommendation generation module recommends satisfactory products to users, and the recommendation evaluation module can continuously monitor and improve the recommendation performance. It not only enhances the user's shopping experience and enables them to discover products of interest and demand more quickly, but also helps companies improve sales efficiency and user loyalty, and achieve sales and profit growth. It can adapt to different application scenarios and market needs, and has a good prospect for promotion and application.

[0095] Figure 4 A schematic diagram of the structure of a product recommendation device provided in an embodiment of the present disclosure.

[0096] like Figure 4 As shown, the product recommendation device 40 includes:

[0097] The first determination module 401 is used to determine the first subject's historical browsing information for each first candidate product, determine the first subject's product preference information for each second candidate product, and determine the first subject's description feedback information for each third candidate product.

[0098] The selection module 402 is used to select a reference candidate product from a plurality of first candidate products according to historical browsing information, select a reference candidate product from a plurality of second candidate products according to product preference information, and select a reference candidate product from a plurality of third candidate products according to description feedback information, wherein the reference candidate product has corresponding candidate product features.

[0099] The second determining module 403 is used to determine the network usage information and attribute information of the first object.

[0100] The third determination module 404 is used to determine the target object group to which the first object belongs according to the network usage information and the attribute information, wherein the target object group has a corresponding group preference feature, and the group preference feature is used to describe the preference of the target object group for the product.

[0101] The recommendation module 405 is used to select at least one target product from some reference candidate products according to the group preference characteristics and the candidate product characteristics.

[0102] It should be noted that the above explanation of the product recommendation method is also applicable to the product recommendation device of this embodiment and will not be repeated here.

[0103] In this embodiment, by determining the historical browsing information of the first object for each first candidate product, determining the product preference information of the first object for each second candidate product, and determining the description feedback information of the first object for each third candidate product, and selecting a reference candidate product from multiple first candidate products according to the historical browsing information, selecting a reference candidate product from multiple second candidate products according to the product preference information, and selecting a reference candidate product from multiple third candidate products according to the description feedback information, wherein the reference candidate product has corresponding candidate product characteristics, determining the network usage information and attribute information of the first object, determining the target object group to which the first object belongs according to the network usage information and attribute information, wherein the target object group has a corresponding group preference feature, the group preference feature is used to describe the preference of the target object group for the product, and selecting at least one target product from some reference candidate products according to the group preference feature and the candidate product feature. Thus, not only can product screening be achieved based on the personalized situation of the first object, but also product screening can be further performed based on the characteristic situation of the target object group to which the first object belongs, so that the screened target product can be more in line with the preference of the first object, and the accuracy of product recommendation can be improved.

[0104] As an example, Figure 5 is a schematic diagram of the structure of an electronic device shown in an exemplary embodiment of the present disclosure, such as Figure 5 As shown, the electronic device 700 may further include:

[0105] The memory 710 and the processor 720, a bus 730 connecting different components (including the memory 710 and the processor 720), the memory 710 stores a computer program, and when the processor 720 executes the program, the product recommendation method of the embodiment of the present disclosure is implemented.

[0106] Bus 730 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus architectures. For example, these architectures include but are not limited to Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus and Peripheral Component Interconnect (PCI) bus.

[0107] The electronic device 700 typically includes a variety of electronic device readable media. These media can be any available media that can be accessed by the electronic device 700, including volatile and non-volatile media, removable and non-removable media.

[0108] The memory 710 may also include computer system readable media in the form of volatile memory, such as random access memory (RAM) 740 and / or cache memory 750. The electronic device 700 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 760 may be used to read and write non-removable, non-volatile magnetic media ( Figure 5 not shown, usually called a "hard drive"). Although Figure 5 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 730 via one or more data medium interfaces. The memory 710 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present disclosure.

[0109] A program / utility 780 having a set (at least one) of program modules 770 may be stored, for example, in the memory 710, such program modules 770 including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. The program modules 770 generally perform the functions and / or methods of the embodiments described in the present disclosure.

[0110] The electronic device 700 may also communicate with one or more external devices 790 (e.g., keyboard, pointing device, display 791, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 700, and / or communicate with any device that enables the electronic device 700 to communicate with one or more other computing devices (e.g., network card, modem, etc.). Such communication may be performed through an input / output (I / O) interface 792. Furthermore, the electronic device 700 may also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN) and / or public network, such as the Internet) through a network adapter 793. As shown, the network adapter 793 communicates with other modules of the electronic device 700 through a bus 730. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 700, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0111] The processor 720 executes various functional applications and data processing by running the programs stored in the memory 710 .

[0112] It should be noted that the implementation process and technical principles of the electronic device of this embodiment refer to the aforementioned explanation of the product recommendation method of the embodiment of the present disclosure, and will not be repeated here.

[0113] In order to implement the above embodiments, the present disclosure also proposes an electronic device, including: a processor, and a memory communicatively connected to the processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory to implement the method provided by the above embodiments.

[0114] In order to implement the above embodiments, the present disclosure further proposes a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the methods provided by the above embodiments.

[0115] In order to implement the above embodiments, the present disclosure also proposes a computer program product, including a computer program, which implements the methods provided by the above embodiments when executed by a processor.

[0116] The collection, storage, use, processing, transmission, provision and disclosure of user personal information involved in this disclosure shall comply with the relevant laws and regulations and shall not violate public order and good morals.

[0117] It should be noted that personal information from users should be collected for legitimate and reasonable purposes and should not be shared or sold outside of these legitimate uses. In addition, such collection / sharing should be carried out after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign the agreement / authorization including authorization of relevant user information before the user uses the function. In addition, any necessary steps should be taken to protect and safeguard access to such personal information data and ensure that others who have access to personal information data comply with its privacy policy and procedures.

[0118] The present disclosure anticipates providing implementation schemes for users to selectively block the use or access of personal information data. That is, the present disclosure anticipates providing hardware and / or software to prevent or block access to such personal information data. Once the personal information data is no longer needed, risks can be minimized by limiting data collection and deleting the data. In addition, when applicable, such personal information is de-identified to protect the privacy of the user.

[0119] In the description of the aforementioned embodiments, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0120] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present disclosure, "plurality" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0121] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present disclosure includes additional implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present disclosure belong.

[0122] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute the instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0123] It should be understood that the various parts of the present disclosure can be implemented in hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used to implement: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0124] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0125] In addition, each functional unit in each embodiment of the present disclosure may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0126] The storage medium mentioned above may be a read-only memory, a disk or an optical disk, etc. Although the embodiments of the present disclosure have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations of the present disclosure. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present disclosure.

Claims

1. A product recommendation method, characterized in that: The following steps are involved: Determine the first subject's historical browsing information for each first candidate product, determine the first subject's product preference information for each second candidate product, and determine the first subject's description feedback information for each third candidate product; Selecting a reference candidate product from a plurality of the first candidate products according to the historical browsing information, selecting a reference candidate product from a plurality of the second candidate products according to the product preference information, and selecting a reference candidate product from a plurality of the third candidate products according to the description feedback information, wherein the reference candidate product has corresponding candidate product features; Determining network usage information and attribute information of the first object; Determining, based on the network usage information and the attribute information, a target object group to which the first object belongs, wherein the target object group has a corresponding group preference feature, and the group preference feature is used to describe the target object group's preference for a product; and At least one target product is selected from some of the reference candidate products according to the group preference characteristics and the candidate product characteristics.

2. The method according to claim 1, characterized in that The determining of the historical browsing information of the first object on each first candidate product includes: Determining the initial browsing information of the first subject for each first candidate product, and identifying abnormal first browsing information from the plurality of the initial browsing information; Analyze the browsing characteristics of the remaining initial browsing information to obtain browsing characteristic information; Correcting the first initial browsing information according to the browsing characteristic information to obtain second browsing information; The remaining initial browsing information and the second browsing information are deduplicated to obtain the historical browsing information.

3. The method according to claim 1, characterized in that The determining the product preference information of the first subject for each second candidate product includes: determining reference product characteristics for each second candidate product; Determining initial preference information of the first subject for each second candidate product, and identifying abnormal first preference information from the plurality of initial preference information, wherein the first preference information has a corresponding second candidate product; Correcting the first preference information according to the reference product features of the second candidate product corresponding to the remaining initial preference information and the reference product features of the second candidate product corresponding to the first preference information to obtain second preference information; The remaining initial preference information and the second preference information are deduplicated to obtain the product preference information.

4. The method according to claim 1, characterized in that The selecting a reference candidate product from the plurality of first candidate products according to the historical browsing information comprises: Determine, based on the historical browsing information, the browsing duration, browsing time, and browsing frequency of each of the first candidate products by the first subject, and determine browsing score information based on the browsing duration, browsing time, and browsing frequency, wherein the browsing score information is positively correlated with the browsing stickiness of the first subject to the first candidate product; A reference candidate product is selected from a plurality of first candidate products according to the browsing score information.

5. The method according to claim 1, characterized in that The selecting a reference candidate product from a plurality of second candidate products according to the product preference information comprises: Determining, based on the product preference information, the purchase frequency, purchase quantity, and purchase time of each of the second candidate products by the first subject; Determining preference score information according to the purchase frequency, purchase quantity and purchase time, wherein the preference score information is positively correlated with the first subject's preference for the second candidate product; A reference candidate product is selected from a plurality of the second candidate products according to the preference score information.

6. The method according to claim 1, characterized in that The selecting a reference candidate product from the plurality of third candidate products according to the description feedback information comprises: Performing semantic recognition on each of the description feedback information to obtain description semantic features; Determining description tendency information corresponding to each of the description semantic features, wherein the description tendency information includes: positive feedback description or negative feedback description; Determining description scoring information according to the positive feedback description or the negative feedback description, wherein the description scoring information is positively correlated with the degree of the positive feedback description; A reference candidate product is selected from the plurality of third candidate products according to the description score information.

7. The method according to claim 1, characterized in that The determining, according to the network usage information and the attribute information, a target object group to which the first object belongs includes: Determine candidate information of at least one candidate object group, wherein the candidate information includes: network usage information and attribute information of each candidate object in the candidate object group; Determining similarities between the network usage information and attribute information and the candidate information, and selecting a maximum similarity from at least one of the similarities; The candidate object group to which the candidate information corresponding to the maximum similarity belongs is taken as the target object group.

8. The method according to claim 1, characterized in that: The selecting at least one target product from part of the reference candidate products according to the group preference feature and the candidate product feature comprises: Inputting the group preference features and the candidate product features into a recommendation model to obtain a recommendation score corresponding to each of the reference candidate products output by the recommendation model, wherein the recommendation model has learned a mapping relationship between the group preference features, the candidate product features, and the recommendation score corresponding to each of the reference candidate products; At least one target product is selected from some of the reference candidate products according to the recommendation score.

9. The method according to claim 1, characterized in that: The method further comprises: Acquire a recommendation description template, wherein the recommendation description template includes: candidate recommendation information corresponding to the candidate preference features and the reference product features; Determine candidate recommendation information corresponding to the group preference feature and the candidate product feature from the recommendation description template, and use the candidate recommendation information corresponding to the group preference feature and the candidate product feature as target recommendation information of the target product; The target recommendation information is provided.

10. A product recommendation device, characterized in that: include: A first determination module is used to determine the first subject's historical browsing information for each first candidate product, determine the first subject's product preference information for each second candidate product, and determine the first subject's description feedback information for each third candidate product; A selection module, configured to select a reference candidate product from the plurality of first candidate products according to the historical browsing information, select a reference candidate product from the plurality of second candidate products according to the product preference information, and select a reference candidate product from the plurality of third candidate products according to the description feedback information, wherein the reference candidate product has corresponding candidate product features; A second determining module, used to determine the network usage information and attribute information of the first object; A third determination module is used to determine the target object group to which the first object belongs according to the network usage information and the attribute information, wherein the target object group has a corresponding group preference feature, and the group preference feature is used to describe the preference of the target object group for a product; and A recommendation module is used to select at least one target product from part of the reference candidate products according to the group preference characteristics and the candidate product characteristics.

11. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 9 when executed by a processor.

13. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 9 when being executed by a processor.

Images