Recommendation content generation method and device, equipment and medium

Abstract

The embodiment of the invention discloses a recommended content generation method and device, equipment and a medium, and the method comprises the steps: obtaining a point-of-interest set in a target scene, the point-of-interest set comprising at least two points of interest and at least one hierarchical label of the points of interest; according to the label similarity between the hierarchical labels of different interest points, updating the hierarchical labels of the interest points to obtain alternative hierarchical labels of the interest points; determining recommended interest points from the interest point set according to at least one of attributes of the interest points, label similarity of alternative hierarchical labels among the interest points and combination quantity of the alternative hierarchical labels; determining recommended subject terms according to the alternative hierarchical labels of the recommended interest points; and generating recommended contents according to the recommended subject terms and the recommended interest points. According to the scheme, the accuracy of the determined recommended interest point is improved, and then the accuracy of the subsequently generated recommended content is improved.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of data processing technology, and in particular to a method, device, equipment and medium for generating recommended content. Background Art

[0002] Recommendation technology has been widely used. In navigation maps, the traditional way is to recommend some POIs (Point of Interest) to users. The selection of these POIs relies on complex recommendation algorithms and AI (Artificial Intelligence) models. However, since the selected algorithms and rules do not consider recommendations based on a theme, the recommended content is often not focused, and users need to understand the recommendation purpose represented by each POI. Therefore, how to accurately generate recommended content is crucial. Summary of the invention

[0003] The present invention provides a method, device, equipment and medium for generating recommended content, so as to improve the accuracy of generated recommended content.

[0004] According to one aspect of the present invention, a method for generating recommended content is provided, comprising:

[0005] Acquire a set of points of interest in a target scene, the set of points of interest comprising at least two points of interest and at least one hierarchical label of the points of interest;

[0006] According to the label similarity between the hierarchical labels of different points of interest, the hierarchical labels of the points of interest are updated to obtain candidate hierarchical labels of the points of interest;

[0007] Determining a recommended point of interest from the set of points of interest according to at least one of the attributes of the points of interest, the tag similarity of the candidate hierarchical tags between the points of interest, and the number of combinations of the candidate hierarchical tags;

[0008] Determining a recommended keyword based on the candidate hierarchical labels of the recommended interest points;

[0009] Generate recommended content based on the recommended keywords and the recommended points of interest.

[0010] According to another aspect of the present invention, there is provided a recommended content generating device, comprising:

[0011] An interest point set acquisition module, used to acquire an interest point set in a target scene, wherein the interest point set includes at least two interest points and at least one hierarchical label of the interest points;

[0012] A candidate hierarchical label determination module is used to update the hierarchical labels of the interest points according to the label similarity between the hierarchical labels of different interest points to obtain candidate hierarchical labels of the interest points;

[0013] a recommended interest point determination module, configured to determine a recommended interest point from the interest point set according to at least one of the attributes of the interest point, the tag similarity of the candidate hierarchical tags between the interest points, and the number of combinations of the candidate hierarchical tags;

[0014] A recommended keyword determining module, used to determine a recommended keyword based on the candidate hierarchical labels of the recommended interest points;

[0015] The recommended content determination module is used to generate recommended content according to the recommended keywords and the recommended interest points.

[0016] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0017] at least one processor; and

[0018] a memory communicatively connected to the at least one processor; wherein,

[0019] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the recommended content generating method described in any embodiment of the present invention.

[0020] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for generating recommended content according to any embodiment of the present invention when executed.

[0021] The embodiment of the present invention provides a solution for generating recommended content, which determines the recommended interest points according to at least one of the attributes of the interest points, the label similarity of the candidate hierarchical labels between the interest points, and the number of combinations of the candidate hierarchical labels, thereby improving the accuracy of the determined recommended interest points, and thus improving the accuracy of the subsequently generated recommended content. At the same time, the embodiment of the present invention determines the candidate hierarchical labels according to the label similarity, thereby avoiding the redundancy of the subsequently generated recommended content due to unnecessary hierarchical labels, and improving the accuracy of the generated recommended content.

[0022] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0024] Figure 1 is a flow chart of a method for generating recommended content provided by Embodiment 1 of the present invention;

[0025] Figure 2 is a flow chart of a method for generating recommended content provided by Embodiment 2 of the present invention;

[0026] Figure 3 is a flow chart of a method for generating recommended content provided by Embodiment 3 of the present invention;

[0027] Figure 4 is a structural diagram of a recommended content generating device provided by a fourth embodiment of the present invention;

[0028] Figure 5 It is a structural diagram of an electronic device for implementing a method for generating recommended content provided by Embodiment 5 of the present invention. DETAILED DESCRIPTION

[0029] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0030] A single-level tag is a tag with only one label, and a hierarchical tag is a tag with more than one label.

[0031] The structure of a hierarchical tag is composed of multiple single-level tags. The multiple single-level tags that constitute a hierarchical tag need to satisfy a certain association relationship in order to form multiple single-level tags into a hierarchical tag. For example: multiple single-level tags appear in a sentence or an attribute unit at the same time, forming a hierarchical tag.

[0032] Both single-level labels and hierarchical labels need to provide multiple vector features, such as: the data source of the label, the number of labels, the rating value of the label, etc. Among them, the data source of the label refers to the label from the name, address, comment, etc., so some types have been predefined when labeling. The number of labels refers to the number of labels in the composition of the label. For example: a single-level label is 1 label, and the number of labels in the hierarchical label is 3, so there are 3 labels. The rating value of the label can refer to the scoring results of each label by the label system.

[0033] Embodiment 1

[0034] Figure 1 This is a flowchart of a method for generating recommended content provided in the first embodiment of the present invention. This embodiment is applicable to the case of generating recommended content according to user needs. The method can be executed by a recommended content generating device, which can be implemented in the form of hardware and / or software. The device can be configured in an electronic device with certain computing capabilities.

[0035] See also Figure 1 The recommended content generation method shown includes:

[0036] S110: Acquire a set of interest points in a target scene, where the set of interest points includes at least two interest points and at least one hierarchical label of the interest points.

[0037] Among them, the target scene refers to the scene where recommended content needs to be generated. Exemplarily, the target scene can be an automatic recommendation scene based on time period and location, and the target scene can also be a user's active query scene. The point of interest set refers to the combination of various points of interest under the target scene. The hierarchical label refers to the identification data associated with the point of interest. Specifically, the hierarchical label can show the key characteristics of the point of interest, that is, the point of interest can be determined based on the hierarchical label. It should be noted that the embodiment of the present invention does not impose any limitation on the number of levels of the hierarchical label of any point of interest, and the number of hierarchical labels. It can be set by technical personnel based on experience. For example, the number of levels of the hierarchical label of any point of interest can be at least two layers (that is, multiple hierarchical labels), and the number of hierarchical labels can be at least one.

[0038] Specifically, at least one level label of at least two points of interest in the target scene is obtained.

[0039] S120 : updating the hierarchical labels of the interest points according to the label similarities between the hierarchical labels of different interest points to obtain candidate hierarchical labels of the interest points.

[0040] Among them, label similarity refers to the degree of association between hierarchical labels used to quantify different points of interest. Alternative hierarchical labels refer to hierarchical labels with a high degree of association between different points of interest. Exemplarily, hierarchical labels with label similarity greater than label values ​​can be used as alternative hierarchical labels. The embodiment of the present invention does not impose any limitation on the size of the label value, which can be set by technicians based on experience.

[0041] For example, if the points of interest include point A and point B, the hierarchical label of point A is a, and the hierarchical label of point B is b, the label similarity between the hierarchical label a and the hierarchical label b is determined; according to the determined label similarity and label value, the hierarchical label a and the hierarchical label b are updated to obtain alternative hierarchical labels for the points of interest.

[0042] S130: Determine a recommended interest point from the interest point set according to at least one of the attributes of the interest point, the tag similarity of the candidate level tags between the interest points, and the number of combinations of the candidate level tags.

[0043] Among them, the attributes of the points of interest refer to the basic attributes of the points of interest. Exemplarily, the attributes of the points of interest may include the source attributes of the points of interest, such as comments, names, and addresses. In the embodiment of the present invention, the similarity of the points of interest can be determined by the similarity of the labels of the same attributes between different points of interest; the similarity of the points of interest is compared with the preset similarity threshold of the points of interest; and according to the comparison result, it is determined whether the corresponding point of interest is a recommended point of interest. Specifically, if the similarity of the points of interest is greater than or equal to the preset similarity threshold of the points of interest, the corresponding point of interest is determined to be a recommended point of interest; if not, the corresponding point of interest is prohibited from being used as a recommended point of interest. Among them, the similarity of the points of interest refers to the degree of association between two points of interest. The embodiment of the present invention does not impose any limitation on the size of the preset point of interest similarity threshold, which can be set by a technician based on experience, or repeatedly determined through a large number of experiments.

[0044] For example, if the attributes of point of interest A are name and address, and the attribute of point of interest B is name, then you can select tags with the same attributes for similarity calculation, such as determining the similarity between the name tag of point of interest A and the name tag of point of interest B, and using the similarity as the point of interest similarity between point of interest A and point of interest B, and determining whether point of interest A and point of interest B are recommended points of interest based on the point of interest similarity.

[0045] Among them, label similarity refers to the degree of association between two candidate hierarchical labels of different points of interest. In an embodiment of the present invention, the label similarities of the candidate hierarchical labels between the points of interest are directly added or weighted to obtain the sum of label similarities; the sum of label similarities is compared with a preset label similarity threshold; based on the comparison result, it is determined whether the corresponding point of interest is a recommended point of interest. Specifically, if the sum of label similarities is greater than or equal to the preset label similarity threshold, the corresponding point of interest is used as a recommended point of interest; if not, the corresponding point of interest is prohibited from being used as a recommended point of interest. The embodiment of the present invention does not impose any limitation on the size of the preset label similarity threshold, which can be set by a technician based on experience, or repeatedly determined through a large number of experiments.

[0046] For example, if the candidate hierarchical labels of point of interest A include a1 and a2, and the candidate hierarchical labels of point of interest B include b1 and b2, where a1 and b1 are a group, and the label similarity of the group is m; a2 and b2 are a group, and the label similarity of the group is n; according to the label similarities m and n, determine whether point of interest A and point of interest B are recommended points of interest. Exemplarily, the label similarities m and n can be directly added or weighted to determine the sum of the label similarities between point of interest A and point of interest B, and the sum of the label similarities is compared with the preset label similarity threshold to determine whether point of interest A and point of interest B are recommended points of interest. Specifically, if the sum of the label similarities is greater than or equal to the preset label similarity threshold, point of interest A and point of interest B are recommended points of interest; if not, point of interest A and point of interest B are prohibited from being recommended points of interest.

[0047] Among them, the number of combinations refers to the number of combinations of alternative hierarchical labels between two points of interest. It should be noted that the embodiment of the present invention does not impose any limitation on the number of combinations, which can be set by technical personnel based on experience. In the embodiment of the present invention, the number of combinations of alternative hierarchical labels between different points of interest is determined; the number of combinations of alternative hierarchical labels between different points of interest is compared with the number of combinations of alternative hierarchical labels set in advance; based on the comparison result, it is determined whether the corresponding point of interest is a recommended point of interest. Specifically, if the number of combinations of alternative hierarchical labels between different points of interest meets the number of combinations of alternative hierarchical labels set in advance, the corresponding point of interest is determined to be a recommended point of interest; if not, the corresponding point of interest is prohibited from being recommended as a point of interest.

[0048] For example, if the number of pre-set combinations of alternative hierarchical labels is greater than or equal to three, the number of combinations of alternative hierarchical labels between interest points A and interest points B is four, the number of combinations of alternative hierarchical labels between interest points A and interest points C is three, and the number of combinations of alternative hierarchical labels between interest points B and interest points C is two, then it can be determined that the number of combinations of alternative hierarchical labels between interest points A and interest points B, and between interest points A and interest points C meets the pre-set number of combinations of alternative hierarchical labels, that is, interest points A, interest points B and interest points C can be determined as recommended interest points.

[0049] It should be noted that, in an embodiment of the present invention, recommended interest points can be determined based on at least one of the attributes of the interest points, the label similarity of the alternative hierarchical labels between the interest points, and the number of combinations of the alternative hierarchical labels. In addition, in an embodiment of the present invention, when using the above-mentioned at least two features to determine the recommended interest points, the order of using at least two features is not limited, and can be set by technicians based on experience.

[0050] S140: Determine recommended keywords based on the candidate hierarchical labels of the recommended interest points.

[0051] Among them, the recommended keywords refer to the key words used to recommend to users.

[0052] S150: Generate recommended content based on the recommended keywords and recommended points of interest.

[0053] Among them, recommended content refers to content recommended to users.

[0054] In an optional embodiment, the recommended keywords and recommended points of interest may be directly used as recommended content.

[0055] In another optional embodiment, in order to improve the richness of the recommended content, the topic description content can be introduced. Exemplarily, generating recommended content according to the recommended topic words and the recommended interest points includes: generating the topic description content according to the recommended topic words; and using the topic description content and the recommended interest points as the recommended content.

[0056] The topic description content refers to a description content generated based on the recommended keywords. The embodiment of the present invention does not limit the method of generating the topic description content, which can be set by technicians based on experience. In an optional embodiment, the recommended keywords can be input into the text macro model to obtain the topic description content.

[0057] For example, if the recommended keywords include bar, drinking, band, music and performance, the recommended keywords are input into the text model to obtain the topic description content. Optionally, the topic description content can be: In this bar full of vitality and music, the crowd is enthusiastic and full of joy. The band's music shuttles through every corner, inspiring people's enthusiasm. The performance on the stage is full of dynamism and vitality, and every musician is fully engaged, immersing people in the ocean of music. This is a place for people to release stress and worries, allowing people to enjoy wonderful music, drink wine happily, forget all worries, and devote themselves to the charm of music. Whether you come here alone or with friends, you can find your own music paradise here and enjoy the dual stimulation of music and alcohol. Or optionally, the topic description content can be: In this vibrant bar, you can enjoy wonderful music and taste mellow drinks with friends. When the drums and guitars sound, the band's performance is immediately full of enthusiasm and passion. Following the rhythm of the music, you can dance your body and release your inner joy. This is a paradise for music lovers. Every performance is a collision of souls, allowing you to immerse yourself in the ocean of music and forget all your troubles. Whether it is a lyrical rock song or a cheerful pop song, this place can satisfy all your desires for music. So, come to this bar and enjoy the charm of music with the band. Let's drink happily and revel until dawn! Or optionally, the theme description content can be: On this exciting night, you can choose to go to a lively bar and enjoy a wonderful music performance. When you step into the door of the bar, the lights are flashing and the music is spreading in the air. You can feel the vitality and enthusiasm of the bar, people are drinking happily and dancing to the rhythm. The band is playing moving music on the stage, and their talent and passion immerse everyone in it. You can sway to the beat of the music, drink happily with friends, and enjoy this energetic and joyful night. Whether you are intoxicated by the frenzy of rock music or intoxicated by the tenderness of jazz, the music performance scene in this bar will bring you an unforgettable experience. Let the combination of music and alcohol ignite the passion in your heart and make this night one of the best memories in your life.

[0058] Furthermore, the topic description content and the recommended interest points are combined to obtain recommended content. Furthermore, the recommended content can be displayed based on a preset display format, wherein the preset display format can be a card format.

[0059] It can be understood that by introducing the topic description content and recommending the interest points and the topic description content as the recommended content, the richness of the recommended content is improved.

[0060] In another optional embodiment, the recommended subject words may be input into the corpus, and corresponding corpora may be matched according to the recommended subject words, and the matched corpora may be used as the subject description content. It should be noted that multiple corpora may be obtained in the corpus according to the recommended subject words, and each corpus is scored, and the corpus with the highest score is selected as the main recommended content.

[0061] The embodiment of the present invention provides a solution for generating recommended content, which determines the recommended interest points according to at least one of the attributes of the interest points, the label similarity of the candidate hierarchical labels between the interest points, and the number of combinations of the candidate hierarchical labels, thereby improving the accuracy of the determined recommended interest points, and thus improving the accuracy of the subsequently generated recommended content. At the same time, the embodiment of the present invention determines the candidate hierarchical labels according to the label similarity, thereby avoiding the redundancy of the subsequently generated recommended content due to unnecessary hierarchical labels, and improving the accuracy of the generated recommended content.

[0062] Embodiment 2

[0063] Figure 2 It is a flowchart of a method for generating recommended content provided by the second embodiment of the present invention. Based on the above embodiments, this embodiment further refines the operation of "updating the hierarchical labels of the interest points according to the label similarity between the hierarchical labels of different interest points to obtain the candidate hierarchical labels of the interest points" into "recording any two interest points in the interest point set as the first interest point and the second interest point respectively; recording the hierarchical label of the first interest point as the first hierarchical label; recording the hierarchical label of the second interest point as the second hierarchical label; generating at least one target label combination corresponding to the interest point set according to the label level of the hierarchical labels, the first hierarchical label and the second hierarchical label; the label level of the first hierarchical label and the second hierarchical label in the target label combination is the same; for each target label combination, determining the label similarity of the target label combination according to the label vector of the hierarchical label in the target label combination; updating the hierarchical labels of the interest points according to the label similarity of at least one target label combination to obtain the candidate hierarchical labels of the interest points", so as to improve the determination mechanism of the candidate hierarchical labels. It should be noted that for the part not described in detail in the embodiment of the present invention, reference can be made to the description of other embodiments.

[0064] See also Figure 2 The recommended content generation method shown includes:

[0065] S210: Acquire a set of interest points in a target scene, where the set of interest points includes at least two interest points and at least one hierarchical label of the interest points.

[0066] S220: Record any two points of interest in the point of interest set as a first point of interest and a second point of interest respectively; record the level label of the first point of interest as a first level label; and record the level label of the second point of interest as a second level label.

[0067] The first point of interest and the second point of interest refer to any two different points of interest in the point of interest set. The first level label refers to the level label corresponding to the first point of interest. The second level label refers to the level label corresponding to the second point of interest.

[0068] S230, generating at least one target tag combination corresponding to the interest point set according to the tag level of the hierarchical tag, the first-level tag and the second-level tag; the tag levels of the first-level tag and the second-level tag in the target tag combination are the same.

[0069] Among them, the label level refers to the number of levels of the hierarchical label. For example, if the hierarchical label is ABCD, the label level is 4. The target label combination refers to the label combination with the same label level of the hierarchical labels between two points of interest. For example, if the first-level label of point of interest x is ABD and the second-level label of point of interest y is AED, then the target label combination is ABD and AED.

[0070] It should be noted that the label levels of the hierarchical labels in the target label combination in the embodiment of the present invention are the same.

[0071] In an optional embodiment, at least one target label combination is generated based on the label level of the hierarchical label, the first-level label and the second-level label, including: combining the first-level label of the first interest point and the second-level label of the second interest point to obtain at least one candidate label combination; for each candidate label combination, if the label levels in the candidate label combination are different, splitting the candidate label combination to obtain the target label combination.

[0072] The candidate tag combination refers to a tag combination obtained by randomly combining the first-level tags and the second-level tags. It should be noted that the candidate tag combination may include tag combinations with different tag levels.

[0073] It can be understood that when the label levels in the candidate label combination are different, the target label combination is obtained by splitting the candidate label combination, thereby achieving the purpose of making the label levels in the target label combination the same, and improving the accuracy of subsequently determining the alternative level labels based on the target label combination.

[0074] In another optional embodiment, for each candidate tag combination, if the tag levels in the candidate tag combination are the same, the candidate tag combination is used as the target tag combination.

[0075] In an optional embodiment, the candidate label combination is split to obtain a target label combination, including: determining a low-level label and a high-level label from the candidate label combination; performing label splitting on the high-level label to obtain at least two candidate low-level labels of the same level as the low-level label; and performing label combination on the low-level label and at least two candidate low-level labels to obtain at least two target label combinations.

[0076] Among them, the low-level label refers to a label with fewer label levels in the hierarchy. The high-level label refers to a label with more label levels in the hierarchy. The candidate low-level label refers to a label with the same label level as the low-level label.

[0077] For example, if the first-level label in the candidate label combination is ABC and the second-level label is DFGH, then the first-level label is a low-level label and the second-level label is a high-level label. The high-level label DFGH is split into C4 3 candidate low-level labels; combine the low-level label ABC with the split candidate low-level labels to obtain the target label combination.

[0078] It can be understood that by introducing candidate low-level labels, the high-level labels are split, the label combinations with different label levels in the candidate label combinations are avoided, and the accuracy of the determined target label combination is improved.

[0079] S240 . For each target tag combination, determine the tag similarity of the target tag combination according to the tag vectors of the hierarchical tags in the target tag combination.

[0080] The label vector refers to the data obtained by vectorizing the hierarchical labels.

[0081] It should be noted that the dimensions of the vector features of all single labels in the embodiment of the present invention are the same. Therefore, when the label levels of the hierarchical labels composed of single labels are the same, the dimensions of the hierarchical labels are also the same. In the embodiment of the present invention, the determination of the label similarity of the target label combination needs to satisfy the characteristics of interchangeability, that is, the label similarity is independent of the order of the single labels of the hierarchical labels in the target label combination. For example, if the hierarchical labels in the target label combination A are ABCD and EFGH, and the label similarity of the target label combination A is Q1; if the target label combination B is BACD and FEGH, and the label similarity of the target label combination B is Q2, then Q1=Q2.

[0082] S250: Update the hierarchical label of the point of interest according to the label similarity of at least one target label combination to obtain a candidate hierarchical label of the point of interest.

[0083] In an optional embodiment, a target tag combination whose tag similarity is greater than a preset tag combination threshold is determined from each target tag combination as a candidate tag combination; based on the candidate tag combination, the hierarchical tag of the point of interest is updated to obtain a candidate hierarchical tag of the point of interest. The candidate hierarchical tag combination refers to a tag combination whose tag similarity is greater than a preset tag combination threshold. The embodiment of the present invention does not impose any limitation on the size of the preset tag combination threshold, which can be set by a technician based on experience, or determined repeatedly through a large number of experiments.

[0084] It should be noted that, for any two different points of interest, if the label similarity of the target label combination of any label level in the target label combination corresponding to the two points of interest is not greater than the preset label combination threshold, the label level is downgraded and the target label combination is updated; the label similarity of the target label combination is re-determined, and the label similarity is compared with the preset label combination threshold; according to the comparison result, the hierarchical label of the point of interest is updated to obtain the candidate hierarchical label of the point of interest. The label level in the embodiment of the present invention is greater than or equal to 2.

[0085] For example, if the label similarities of the target label combinations with a label level of 3 are not greater than the preset label combination threshold, the label level is reduced to 2, the target label combination is updated, the label similarity of the target label combination is re-determined, and the label similarity is compared with the preset label combination threshold; based on the comparison results, the hierarchical labels of the points of interest are updated to obtain alternative hierarchical labels for the points of interest.

[0086] S260: Determine a recommended interest point from the interest point set according to at least one of the attributes of the interest point, the tag similarity of the candidate level tags between the interest points, and the number of combinations of the candidate level tags.

[0087] S270: Determine recommended keywords based on the candidate hierarchical labels of the recommended interest points.

[0088] S280: Generate recommended content based on the recommended keywords and the recommended points of interest.

[0089] A method for generating recommended content provided by an embodiment of the present invention improves the accuracy of the determined label similarity by determining the label similarity based on the label vector; at the same time, by introducing a target label combination so that the label levels of the hierarchical labels between different interest points are the same, the accuracy of the subsequently determined alternative hierarchical labels is improved.

[0090] Embodiment 3

[0091] Figure 3 It is a flowchart of a method for generating recommended content provided in the third embodiment of the present invention. Based on the above embodiments, the embodiment of the present invention further refines the operation of "determining recommended interest points from a set of interest points based on at least one of the attributes of the interest points, the label similarity of the alternative hierarchical labels between the interest points, and the number of combinations of the alternative hierarchical labels" into "determining at least two candidate interest points from a set of interest points based on the attributes of the interest points, the label similarity of the alternative hierarchical labels between the interest points, and the number of combinations of the alternative hierarchical labels; determining recommended interest points from at least two candidate interest points based on the user's portrait label and the alternative hierarchical labels of the candidate interest points" to improve the determination mechanism of recommended interest points. It should be noted that for the parts not described in detail in the embodiments of the present invention, reference may be made to the descriptions of other embodiments.

[0092] See also Figure 3 The recommended content generation method shown includes:

[0093] S310: Acquire a set of interest points in a target scene, where the set of interest points includes at least two interest points and at least one hierarchical label of the interest points.

[0094] S320: updating the hierarchical labels of the interest points according to the label similarities between the hierarchical labels of different interest points to obtain candidate hierarchical labels of the interest points.

[0095] S330: Determine at least two candidate interest points from the interest point set according to at least one of the attributes of the interest points, the label similarity of the candidate level labels between the interest points, and the number of combinations of the candidate level labels.

[0096] Among them, the candidate interest point refers to the interest point corresponding to the alternative level label that satisfies at least one of the attributes of the interest point, the label similarity of the alternative level labels between the interest points, and the number of combinations of the alternative level labels in the interest point set.

[0097] S340: Determine a recommended point of interest from at least two candidate points of interest according to the user's portrait label and the candidate level labels of the candidate points of interest.

[0098] Among them, the portrait tag can be used to characterize the needs or habits of the user. Exemplarily, the portrait tag can include a portrait single tag and a portrait hierarchical tag. Among them, the portrait single tag refers to a single tag, that is, the tag level of the portrait tag is 1. The portrait hierarchical tag refers to multiple tags, that is, the tag level of the portrait hierarchical tag is at least 2.

[0099] In an optional embodiment, a recommended interest point is determined from at least two candidate interest points based on the user's portrait tag and the alternative level tags of the candidate interest points, including: determining the portrait similarity between the user's portrait tag and the alternative level tags of the candidate interest points; and determining the recommended interest point from at least two candidate interest points based on the portrait similarity.

[0100] The portrait similarity is used to quantify the similarity between the portrait label and the candidate level label. Exemplarily, the portrait similarity can be determined according to the method for determining label similarity.

[0101] In an embodiment of the present invention, each candidate hierarchical label corresponding to any candidate point of interest is determined; the portrait similarity between each candidate hierarchical label and the portrait label is determined respectively; at least one portrait similarity of the candidate point of interest is directly summed or weightedly summed to obtain the sum of portrait similarities; the sum of portrait similarities is compared with a preset portrait similarity threshold; and based on the comparison result, a recommended point of interest is determined. Specifically, if the sum of portrait similarities is greater than or equal to the preset portrait similarity threshold, the corresponding candidate point of interest is used as a recommended point of interest; if not, the candidate point of interest is prohibited from being used as a recommended point of interest. It should be noted that the embodiment of the present invention does not impose any limitation on the size of the preset portrait similarity threshold, which can be set by a technician based on experience, or determined repeatedly through a large number of experiments.

[0102] It should be noted that in the embodiment of the present invention, after determining the portrait similarity, the candidate level labels corresponding to the portrait similarity can be screened according to the preset portrait screening threshold; the candidate level labels greater than or equal to the preset portrait screening threshold are retained; the candidate level labels less than the preset portrait screening threshold are deleted; and then the recommended interest points are determined according to the sum of the portrait similarities corresponding to the retained candidate level labels and the preset portrait similarity threshold. The advantage of this is that it realizes the initial screening of the candidate level labels according to the portrait similarity, improves the efficiency of subsequently determining the recommended interest points from at least two candidate interest points according to the portrait similarity, and reduces the waste of resources.

[0103] It can be understood that by determining the recommended interest points based on the portrait similarity, the accuracy of the determined recommended interest points is improved.

[0104] In an optional embodiment, determining the portrait similarity between the user's portrait tag and the alternative hierarchical tags of the candidate interest point includes: if the portrait hierarchical tag is at least two single tags, combining the at least two single tags to obtain the portrait hierarchical tag; generating at least one portrait tag combination based on the portrait hierarchical tag and the alternative hierarchical tags of the candidate interest point; the label levels between the portrait hierarchical tags and the alternative hierarchical tags in the portrait tag combination are the same; and determining the portrait similarity between the portrait hierarchical tags and the alternative hierarchical tags in the portrait tag combination.

[0105] Among them, the portrait tag combination refers to the combination of a portrait-level tag and an alternative-level tag when the tag levels are the same. The tag level refers to the number of levels of the level tags. Correspondingly, the same tag level means that the number of tag levels of the portrait-level tag in the portrait tag combination is the same as the number of tag levels of the alternative-level tag.

[0106] Optionally, at least one image label combination is generated based on the image level label and the candidate level label of the candidate interest point, including: label combination of the image level label and the candidate level label of the candidate interest point to obtain at least one reference label combination; for each reference label combination, if the label levels in the reference label combination are different, the reference label combination is split to obtain an image label combination; if the label levels in the reference label combination are the same, the reference label combination is directly used as the image label combination. Wherein, the reference label combination refers to a combination consisting of an image level label and an alternative level label without considering the label level.

[0107] The different label levels refer to that the number of label levels of the portrait-level label and the number of label levels of the alternative-level label in the portrait label combination are different.

[0108] Specifically, the reference label combination is split to obtain a portrait label combination, including determining a low-level label and a high-level label from the portrait-level label or the candidate-level labels of the candidate interest points; performing label splitting on the high-level label to obtain at least two candidate low-level labels of the same level as the low-level label; and performing label combination on the low-level label and at least two candidate low-level labels to obtain at least two portrait label combinations.

[0109] In an embodiment of the present invention, the portrait similarity between the user's portrait tag and the candidate level tag of the candidate interest point is determined, including: if the label levels between the portrait level tag and the candidate level tag of the candidate interest point are different, adjusting the portrait level tag or the candidate level tag of the candidate interest point; determining the portrait similarity between the portrait level tag and the adjusted alternative level tag of the candidate interest point.

[0110] It can be understood that by determining the portrait similarity based on the portrait level labels and the alternative level labels, the accuracy of the determined portrait similarity is improved; at the same time, by introducing the portrait level labels, the omission of portrait labels is avoided and the accuracy of the determined portrait label combination is improved.

[0111] In another optional embodiment, if the portrait label is a single label, at least one portrait label combination is generated directly based on the portrait label and the alternative level labels of the candidate interest point; the label levels between the portrait level label and the alternative level labels in the portrait label combination are the same; and the portrait similarity between the portrait level label and the alternative level labels in the portrait label combination is determined.

[0112] S350: Determine a recommended keyword based on the candidate hierarchical labels of the recommended interest point.

[0113] In an optional embodiment, determining recommended keywords based on candidate hierarchical labels of recommended interest points includes: combining portrait labels whose portrait similarity is greater than a similarity threshold as a keyword label combination; and determining recommended keywords based on the keyword label combination.

[0114] The theme word label combination refers to a portrait label combination whose portrait similarity is greater than a similarity threshold. Exemplarily, the theme word label combination may include a portrait level label and an alternative level label for a recommended point of interest. The embodiment of the present invention does not impose any limitation on the size of the similarity threshold, which may be set by a technician based on experience or determined repeatedly through a large number of experiments.

[0115] It can be understood that by introducing a similarity threshold and screening the portrait label combinations, the influence of irrelevant portrait label combinations on the determination of recommended keywords is avoided, the waste of resources is reduced, the accuracy of the determined keyword label combinations is improved, and the accuracy of the determined recommended keywords is further improved.

[0116] In an optional embodiment, the method further includes: sorting the recommended subject words according to the tag similarities of the subject word tag combinations.

[0117] For example, if the topic word label combinations include: ABCD and ACDF combination, EFA and ACF combination, and AFD and AFE combination; among them, the label similarity of ABCD and ACDF combination is 0.8; the label similarity of EFA and ACF combination is 0.9; the label similarity of AFD and AFE is 0.85, then according to the label similarity of each combination, the occurrence probability of a single label is determined, as shown in Table 1:

[0118] Specifically, according to the data in Table 1, it can be determined that the occurrence probability of label A is 5.1, the occurrence probability of label B is 0.8, the occurrence probability of label C is 2.5, the occurrence probability of label D is 2.45, the occurrence probability of label E is 1.75, and the occurrence probability of label F is 4.3; according to the occurrence probability of each label, each label is sorted, that is, the recommended keywords are sorted from large to small according to the occurrence probability, and the recommended keywords are obtained: label A, label F, label C, label D, label E, label B.

[0119] Table 1 Occurrence probability of a single label

[0120] Keyword tag combination A B C D E F ABCD 0.8 0.8 0.8 0.8 0 0 ACDF 0.8 0 0.8 0.8 0 0.8 EFA 0.9 0 0 0 0.9 0.9 ACF 0.9 0 0.9 0 0 0.9 AFD 0.85 0 0 0.85 0 0.85 AFE 0.85 0 0 0 0.85 0.85 total 5.1 0.8 2.5 2.45 1.75 4.3

[0121] It can be understood that by sorting the recommended keywords according to the label similarity of the keyword label combination, adjusting the process of generating recommended content based on the recommended keywords and recommended interest points, the recommended keywords ranked in front can be prioritized as keywords for the recommended content according to the input sorting, thereby achieving better control over the generation of recommended content, improving the accuracy of the subsequently generated recommended content, and achieving better highlighting of the theme.

[0122] S360: Generate recommended content based on the recommended keywords and the recommended points of interest.

[0123] The recommended content generation scheme provided by the embodiment of the present invention determines at least two candidate interest points from a set of interest points based on the attributes of the interest points, the label similarity of the alternative hierarchical labels between the interest points, and at least one of the number of combinations of the alternative hierarchical labels, thereby determining the candidate interest points from multiple dimensions and improving the accuracy of the determined candidate interest points. At the same time, based on the user's portrait label and the alternative hierarchical labels of the candidate interest points, the recommended interest point is determined from at least two candidate interest points, thereby determining the recommended interest point on the basis of considering the user's needs and improving the accuracy of the determined recommended interest point.

[0124] Embodiment 4

[0125] Figure 4 1 is a schematic diagram of the structure of a recommended content generation device provided by Embodiment 4 of the present invention. This embodiment is applicable to the case of generating recommended content according to user needs. The method can be executed by a recommended content generation device, which can be implemented in the form of hardware and / or software, and can be configured in an electronic device with certain computing capabilities.

[0126] like Figure 4As shown, the device includes: an interest point set acquisition module 410, an alternative level label determination module 420, a recommended interest point determination module 430, a recommended keyword determination module 440 and a recommended content determination module 450.

[0127] An interest point set acquisition module 410 is used to acquire an interest point set in a target scene, wherein the interest point set includes at least two interest points and at least one hierarchical label of the interest points;

[0128] A candidate hierarchical label determination module 420 is used to update the hierarchical label of the interest point according to the label similarity between the hierarchical labels of different interest points to obtain the candidate hierarchical label of the interest point;

[0129] a recommended interest point determination module 430, configured to determine a recommended interest point from the interest point set according to at least one of the attributes of the interest point, the label similarity of the candidate hierarchical labels between the interest points, and the number of combinations of the candidate hierarchical labels;

[0130] A recommended keyword determining module 440 is used to determine a recommended keyword based on the candidate hierarchical labels of the recommended interest points;

[0131] The recommended content determination module 450 is used to generate recommended content according to the recommended keywords and the recommended interest points.

[0132] The embodiment of the present invention provides a solution for generating recommended content, which determines the recommended interest points according to at least one of the attributes of the interest points, the label similarity of the candidate hierarchical labels between the interest points, and the number of combinations of the candidate hierarchical labels, thereby improving the accuracy of the determined recommended interest points, and thus improving the accuracy of the subsequently generated recommended content. At the same time, the embodiment of the present invention determines the candidate hierarchical labels according to the label similarity, thereby avoiding the redundancy of the subsequently generated recommended content due to unnecessary hierarchical labels, and improving the accuracy of the generated recommended content.

[0133] Optionally, the candidate level label determination module 420 includes:

[0134] An interest point determination unit, configured to record any two interest points in the interest point set as a first interest point and a second interest point respectively; the hierarchical label of the first interest point is recorded as a first hierarchical label; the hierarchical label of the second interest point is recorded as a second hierarchical label;

[0135] A target tag combination generating unit, configured to generate at least one target tag combination corresponding to the interest point set according to the tag level of the hierarchical tag, the first hierarchical tag and the second hierarchical tag; the tag levels of the first hierarchical tag and the second hierarchical tag in the target tag combination are the same;

[0136] A label similarity determination unit, configured to determine, for each target label combination, the label similarity of the target label combination according to the label vectors of the hierarchical labels in the target label combination;

[0137] The candidate hierarchical label determination unit is used to update the hierarchical label of the interest point according to the label similarity of at least one target label combination to obtain the candidate hierarchical label of the interest point.

[0138] Optionally, a target label combination generating unit includes:

[0139] a candidate tag combination determining subunit, configured to perform tag combination on the first level tag of the first interest point and the second level tag of the second interest point to obtain at least one candidate tag combination;

[0140] The target label combination determination subunit is used to split each candidate label combination, if the label levels in the candidate label combination are different, to obtain a target label combination.

[0141] Optionally, the target tag combination determination unit is specifically used to:

[0142] Determine a low-level label and a high-level label from the candidate label combination;

[0143] Performing label splitting on the high-level label to obtain at least two candidate low-level labels of the same level as the low-level label;

[0144] The low-level label and at least two candidate low-level labels are combined to obtain at least two target label combinations.

[0145] Optionally, the recommended interest point determination module 430 includes:

[0146] a candidate interest point determination unit, configured to determine at least two candidate interest points from the interest point set according to at least one of the attributes of the interest points, the label similarity of the candidate hierarchical labels between the interest points, and the number of combinations of the candidate hierarchical labels;

[0147] The recommended interest point determination unit is used to determine a recommended interest point from the at least two candidate interest points according to the user's portrait label and the candidate level labels of the candidate interest points.

[0148] Optionally, the recommended interest point determination unit includes:

[0149] A portrait similarity determination subunit, used to determine the portrait similarity between the user's portrait tag and the candidate level tag of the candidate interest point;

[0150] The recommended interest point determination subunit is used to determine a recommended interest point from the at least two candidate interest points according to the portrait similarity.

[0151] Optionally, the portrait similarity determination subunit is specifically used for:

[0152] If the portrait-level label is at least two single labels, the at least two single labels are combined to obtain a portrait-level label;

[0153] generating at least one image label combination according to the image level label and the candidate level label of the candidate interest point; wherein the image level label and the candidate level label in the image label combination have the same label level;

[0154] Determine the image similarity between the image level label and the candidate level label in the image label combination.

[0155] Optionally, the recommended keyword determination module 440 includes:

[0156] A theme label combination determination unit, used to use the image label combination whose image similarity is greater than a similarity threshold as the theme label combination;

[0157] The recommended keyword determining unit is used to determine the recommended keyword according to the keyword tag combination.

[0158] Optionally, the device further comprises:

[0159] The recommended word ranking unit is used to rank the recommended subject words according to the tag similarity of the subject word tag combination.

[0160] Optionally, the recommended content determination module 450 includes:

[0161] A subject description content generating unit, used to generate subject description content according to the recommended subject words;

[0162] The recommended content determination unit is used to use the subject description content and the recommended interest points as recommended content.

[0163] The recommended content generating device provided in the embodiment of the present invention can execute the recommended content generating method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing each recommended content generating method.

[0164] In the technical solution of the present invention, the collection, storage, use, processing, transmission, provision and disclosure of data such as the point of interest set, the attributes of the point of interest and the label similarity are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0165] Embodiment 5

[0166] Figure 5 5 is a schematic diagram of the structure of an electronic device for implementing a method for generating recommended content provided by Embodiment 5 of the present invention. Electronic device 510 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The switching device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0167] like Figure 5 As shown, the switching device 510 includes at least one processor 511, and a memory connected to the at least one processor 511, such as a read-only memory (ROM) 512, a random access memory (RAM) 513, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 511 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 512 or the computer program loaded from the storage unit 518 to the random access memory (RAM) 513. In the RAM 513, various programs and data required for the operation of the switching device 510 can also be stored. The processor 511, the ROM 512, and the RAM 513 are connected to each other via a bus 514. An input / output (I / O) interface 515 is also connected to the bus 514.

[0168] A number of components in the switching device 510 are connected to the I / O interface 515, including: an input unit 516, such as a keyboard, a mouse, etc.; an output unit 517, such as various types of displays, speakers, etc.; a storage unit 518, such as a disk, an optical disk, etc.; and a communication unit 519, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 519 allows the switching device 510 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0169] The processor 511 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 511 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 511 performs the various methods and processes described above, such as a recommended content generation method.

[0170] In some embodiments, the recommended content generation method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 518. In some embodiments, part or all of the computer program may be loaded and / or installed on the switching device 510 via the ROM 512 and / or the communication unit 519. When the computer program is loaded into the RAM 513 and executed by the processor 511, one or more steps of the recommended content generation method described above may be performed. Alternatively, in other embodiments, the processor 511 may be configured to execute the recommended content generation method in any other appropriate manner (e.g., by means of firmware).

[0171] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0172] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0173] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0174] To provide interaction with a user, the systems and techniques described herein may be implemented on a switching device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the switching device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0175] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0176] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0177] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0178] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for generating recommended content, characterized in that: include: Acquire a set of points of interest in a target scene, the set of points of interest comprising at least two points of interest and at least one hierarchical label of the points of interest; According to the label similarity between the hierarchical labels of different points of interest, the hierarchical labels of the points of interest are updated to obtain candidate hierarchical labels of the points of interest; Determining a recommended point of interest from the set of points of interest according to at least one of the attributes of the points of interest, the tag similarity of the candidate hierarchical tags between the points of interest, and the number of combinations of the candidate hierarchical tags; Determining a recommended keyword based on the candidate hierarchical labels of the recommended interest points; Generate recommended content based on the recommended keywords and the recommended points of interest.

2. The method according to claim 1, characterized in that According to the label similarity between the hierarchical labels of different points of interest, the hierarchical labels of the points of interest are updated to obtain candidate hierarchical labels of the points of interest, including: Record any two points of interest in the set of points of interest as a first point of interest and a second point of interest respectively; record the level label of the first point of interest as a first level label; record the level label of the second point of interest as a second level label; Generate at least one target tag combination corresponding to the interest point set according to the tag level of the hierarchical tag, the first-level tag and the second-level tag; the tag levels of the first-level tag and the second-level tag in the target tag combination are the same; For each target label combination, determine the label similarity of the target label combination according to the label vectors of the hierarchical labels in the target label combination; According to the label similarity of at least one target label combination, the hierarchical label of the point of interest is updated to obtain a candidate hierarchical label of the point of interest.

3. The method according to claim 2, characterized in that At least one target tag combination is generated according to the tag level of the hierarchical tag, the first-level tag, and the second-level tag, including: Performing a label combination on the first level label of the first interest point and the second level label of the second interest point to obtain at least one candidate label combination; For each candidate label combination, if the label levels in the candidate label combination are different, the candidate label combination is split to obtain the target label combination.

4. The method according to claim 3, characterized in that The candidate label combination is split to obtain the target label combination, including: Determine a low-level label and a high-level label from the candidate label combination; Performing label splitting on the high-level label to obtain at least two candidate low-level labels of the same level as the low-level label; The low-level label and at least two candidate low-level labels are combined to obtain at least two target label combinations.

5. The method according to claim 1, characterized in that: Determining a recommended point of interest from the set of points of interest according to at least one of the attributes of the points of interest, the tag similarity of the candidate hierarchical tags between the points of interest, and the number of combinations of the candidate hierarchical tags, comprises: Determining at least two candidate interest points from the set of interest points according to at least one of the attributes of the interest points, the label similarity of the candidate hierarchical labels between the interest points, and the number of combinations of the candidate hierarchical labels; Determine a recommended point of interest from the at least two candidate points of interest according to the user's portrait tag and the candidate level tags of the candidate points of interest.

6. The method according to claim 5, characterized in that Determining a recommended point of interest from the at least two candidate points of interest according to the user's portrait tag and the candidate level tags of the candidate points of interest includes: Determining the similarity between the user's portrait label and the candidate hierarchical labels of the candidate interest points; According to the portrait similarity, a recommended interest point is determined from the at least two candidate interest points.

7. The method according to claim 6, characterized in that Determining the similarity between the user's portrait label and the candidate level label of the candidate interest point includes: If the portrait label is at least two single labels, the at least two single labels are combined to obtain a portrait level label; generating at least one image label combination according to the image level label and the candidate level label of the candidate interest point; wherein the image level label and the candidate level label in the image label combination have the same label level; Determine the image similarity between the image level label and the candidate level label in the image label combination.

8. The method according to claim 7, characterized in that: Determining a recommended keyword according to the candidate hierarchical labels of the recommended interest point includes: The image label combination whose image similarity is greater than the similarity threshold is used as the theme label combination; According to the keyword tag combination, a recommended keyword is determined.

9. The method according to claim 8, characterized in that Also includes: The recommended keywords are sorted according to the tag similarities of the keyword tag combinations.

10. The method according to claim 1, characterized in that Generating recommended content according to the recommended keywords and the recommended points of interest includes: Generating topic description content according to the recommended topic words; The subject description content and the recommended interest points are used as recommended content.

11. A recommended content generating device, characterized in that: include: An interest point set acquisition module, used to acquire an interest point set in a target scene, wherein the interest point set includes at least two interest points and at least one hierarchical label of the interest points; A candidate hierarchical label determination module is used to update the hierarchical labels of the interest points according to the label similarity between the hierarchical labels of different interest points to obtain candidate hierarchical labels of the interest points; a recommended interest point determination module, configured to determine a recommended interest point from the interest point set according to at least one of the attributes of the interest point, the tag similarity of the candidate hierarchical tags between the interest points, and the number of combinations of the candidate hierarchical tags; A recommended keyword determining module, used to determine a recommended keyword based on the candidate hierarchical labels of the recommended interest points; The recommended content determination module is used to generate recommended content according to the recommended keywords and the recommended interest points.

12. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the recommended content generating method according to any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the recommended content generation method according to any one of claims 1 to 10 when executed.

Images