An intelligent video recommendation method, device and storage medium

Abstract

This application provides an intelligent video recommendation method, device, and storage medium, relating to the field of video recommendation technology. It includes the following steps: S1, labeling videos with multiple tag groups; S2, constructing an explicit tag preference library for users, storing the user's preference for each explicit tag; S3, updating the user's preference for each explicit tag in response to the user's historical video browsing history; S4, combining the current popularity of the video to be recommended with the product of the user's preference for the explicit tags of the video to be recommended and a weighting factor to obtain the priority value of the video to be recommended; S5, sorting the videos to be recommended according to their priority values ​​from largest to smallest, and selecting the highest-priority video for recommendation. This application, in the video recommendation process, takes into account both the user's preference for events and different public opinions on the same event, avoiding the phenomenon of information cocoons caused by excessive focus on a single viewpoint.

Description

Technical Field

[0001] This application relates to the field of video recommendation technology, and in particular to an intelligent video recommendation method, device and storage medium. Background Technology

[0002] "Information cocoon" refers to the phenomenon where people become obsessed with a particular interest in information, thus limiting and confining their understanding to a "cocoon" like a silkworm cocoon, resulting in ignorance of other aspects they are not interested in.

[0003] In today's era of short video popularity, most video recommendation platforms use algorithms that analyze users' historical video viewing records to find topics that users are interested in and recommend corresponding videos, which easily leads to the phenomenon of information cocoons.

[0004] Existing technologies have also proposed some solutions to avoid information cocoons. For example, Chinese invention patent application CN117150081A proposes a method, device, electronic device and storage medium for video recommendation. Based on the user's own interests, it gives a certain degree of emphasis to videos that the user has not played, thereby ensuring that the recommended videos not only match the user's interests but also have a certain degree of novelty, effectively avoiding the information cocoon effect and thus improving the user experience.

[0005] However, breaking the information cocoon isn't just about recommending more points of interest, such as videos containing specific times or exposing social phenomena. More importantly, video creators often express their own opinions on the event, which are frequently subjective and biased. If video recommendation methods only focus on recommending videos containing more events, they risk falling into another trap: users receive only the same or similar viewpoints on the same event, ultimately affecting society's perception and judgment of that particular event.

[0006] Therefore, devising a method that can balance different perspectives on the same event in a video is one of the pressing problems that needs to be solved. Summary of the Invention

[0007] This application provides an intelligent video recommendation method, device, and storage medium to at least solve the above-mentioned technical problems existing in the prior art.

[0008] According to a first aspect of this application, an intelligent video recommendation method is provided, comprising the following steps:

[0009] S1, which is a multi-tag group for video annotation; the multi-tag group includes explicit tags and implicit tags;

[0010] S2, Build an explicit tag preference library for the user, wherein the explicit tag preference library stores the user's preference for each explicit tag;

[0011] S3, responding to the user's historical video browsing history, updates the user's preference for each explicit tag stored in the explicit tag preference library;

[0012] S4. The priority value of the video to be recommended is obtained by combining the current popularity of the video to be recommended with the product of the user's preference for the explicit tags of the video to be recommended and the weight factor. The weight factor is negatively correlated with the number of times the implicit tags of the video to be recommended have been viewed in the historical video browsing records.

[0013] S5 sorts the videos to be recommended according to their priority values ​​from highest to lowest, and selects the videos with the highest priority to be recommended.

[0014] In some embodiments of the first aspect of this application, in step S2, when constructing the explicit tag preference library, an initial preference degree is assigned to each explicit tag; and each time a new explicit tag is generated, the new explicit tag is updated to the explicit tag preference library, and an initial preference degree is assigned at the same time.

[0015] In some embodiments of the first aspect of this application, in step S3, the method for updating the user's preference for each explicit tag stored in the explicit tag preference library in response to the user's historical video browsing records is as follows:

[0016] S301, Set a first time threshold and a second time threshold, wherein the first time threshold is less than the second time threshold;

[0017] S302, Extract the user's historical video browsing records within a specified time period;

[0018] S303, Statistically analyze the browsing time and explicit tags in the multiple tag groups of each video in the historical video browsing records;

[0019] S304, if the video viewing time is greater than the first time threshold, then increase the user's preference for those explicit tags in the explicit tag preference library; if the video viewing time is less than the first time threshold, then decrease the user's preference for those explicit tags in the explicit tag preference library; if the video viewing time is between the first time threshold and the second time threshold, then keep the user's preference for those explicit tags unchanged in the explicit tag preference library.

[0020] In some embodiments of the first aspect of this application, in step S3, the method for updating the user's preference for each explicit tag stored in the explicit tag preference library in response to the user's historical video browsing records is as follows:

[0021] S310, each video is configured with a feedback mechanism, which is triggered during or after video playback to display explicit tags of the video to the user for positive or negative feedback.

[0022] S311, Set a first time threshold and a second time threshold, wherein the first time threshold is less than the second time threshold;

[0023] S312, Extract the user's historical video browsing records within a specified time period;

[0024] S313, Statistically analyze the browsing time of each video in the historical video browsing records, the explicit tags in the video's multiple tag groups, and the user's feedback on the explicit tags.

[0025] S314, if the video viewing time is greater than the first time threshold, then within the explicit tag preference library, increase the user's first preference for these explicit tags; if the video viewing time is less than the first time threshold, then within the explicit tag preference library, decrease the user's second preference for these explicit tags; if the video viewing time is between the first time threshold and the second time threshold, then within the explicit tag preference library, keep the user's preference for these explicit tags unchanged.

[0026] S315, if the user provides positive feedback on the explicit labels, the user's third preference for those explicit labels increases; conversely, if the user provides negative feedback on the explicit labels, the user's fourth preference for those explicit labels decreases; if the user does not provide feedback on the explicit labels, the user's preference for those explicit labels remains unchanged.

[0027] In some embodiments of the first aspect of this application, the method for obtaining the priority value of the video to be recommended in step S4 by combining the current popularity of the video to be recommended with the product of the user's preference for the explicit tags of the video to be recommended and a weight factor is as follows:

[0028] S401, construct weight factors that are negatively correlated with the implicit tags of the video to be recommended;

[0029] S402, calculate the current popularity of the video to be recommended;

[0030] S403, calculate the current popularity of the video to be recommended, and the sum of the product of the user's preference for the explicit tags of the video to be recommended and the weight factor, as the priority value for obtaining the video to be recommended.

[0031] In some embodiments of the first aspect of this application, the method for constructing weight factors that are negatively correlated with the implicit tags of the video to be recommended in step S401 is as follows:

[0032] Extract the user's historical video browsing history within a specified time period, and count the videos that contain the explicit tags corresponding to the implicit tags of the videos to be recommended, and record them as the first video;

[0033] The total number of first videos and the number of first videos containing the implicit tags of the video to be recommended are counted.

[0034] Calculate the ratio of the number of first videos containing the implicit tag of the video to be recommended to the total number of first videos, and subtract this ratio as a constant to obtain the weighting factor.

[0035] In some embodiments of the first aspect of this application, the method for calculating the current popularity of the video to be recommended in step S402 is as follows:

[0036] Extract the explicit tags of the video to be recommended, and count the number of times all users have viewed the video to be recommended within the specified time period, which is recorded as the first time count;

[0037] Divide the first number of times by the set number threshold to obtain the current popularity of the video to be recommended.

[0038] In some embodiments of the first aspect of this application, the determination of whether a user has viewed the video to be recommended is based on the following criteria: when the user's viewing time of the video to be recommended exceeds a first time threshold, and / or the user provides positive feedback on the video to be recommended, the user is deemed to have viewed the video to be recommended.

[0039] In some embodiments of the first aspect of this application, if the number of times a user views a video marked with a certain explicit tag for a period of time exceeding a first time threshold reaches a set number threshold, and / or the number of times a user provides positive feedback on a certain explicit tag reaches a set number threshold, the user will be marked as a user followed by that explicit tag.

[0040] In S1, verified video publishers are given corresponding explicit tags and notification tags for reporting events. In S5, when prioritizing videos, if one of the videos to be recommended is labeled with a notification tag, the video labeled with the notification tag will be given the highest priority and ranked first when recommended to users who follow the corresponding explicit tag.

[0041] According to a second aspect of this application, an intelligent video recommendation system is provided, comprising:

[0042] The tagging module is used to annotate videos with multiple tag groups; the multiple tag groups include explicit tags and implicit tags;

[0043] The preference building module builds an explicit tag preference library for users, which stores the user's preference for each explicit tag.

[0044] The preference update module responds to the user's historical video browsing history and updates the user's preference for each explicit tag stored in the explicit tag preference library;

[0045] The priority calculation module combines the current popularity of the video to be recommended with the product of the user's preference for the explicit tags of the video to be recommended and the weight factor to obtain the priority value of the video to be recommended. The weight factor is negatively correlated with the number of times the implicit tags of the video to be recommended have been viewed in the historical video browsing records.

[0046] The sorting and recommendation module sorts the videos to be recommended according to their priority values ​​from highest to lowest, and selects the first number of videos with the highest priority to recommend.

[0047] According to a third aspect of this application, an electronic device is provided, comprising:

[0048] At least one processor; and

[0049] A memory communicatively connected to the at least one processor; wherein,

[0050] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method described in this application.

[0051] According to a fourth aspect of this application, a non-transitory computer-readable storage medium is provided storing computer instructions for causing the computer to perform the methods described in this application.

[0052] Compared with the prior art, this application has the following advantages:

[0053] This application employs a multi-tag grouping method for video labeling. Firstly, it combines users' historical video browsing records to update their preference for each explicit tag, thus prioritizing the recommendation of videos with higher-preference explicit tags in subsequent videos. Secondly, it uses a weighted summation based on a weighting factor, where the weighting factor is negatively correlated with the number of times implicit tags are viewed in historical video browsing records. For videos labeled with the same explicit tag, those with different implicit tags are prioritized for recommendation. Therefore, in the video recommendation process, it considers both users' preferences for events and different public opinions on the same event, avoiding the information cocoon phenomenon caused by excessive focus on a single viewpoint.

[0054] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description

[0055] The above and other objects, features, and advantages of exemplary embodiments of this application will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings. Several embodiments of this application are illustrated in the drawings by way of example and not limitation, wherein:

[0056] In the accompanying drawings, the same or corresponding reference numerals indicate the same or corresponding parts.

[0057] Figure 1 The overall flowchart of the intelligent video recommendation method of this application is shown.

[0058] Figure 2 The multi-label group architecture diagram of this application is shown.

[0059] Figure 3 A flowchart of one of the explicit tag preference update methods of this application is shown.

[0060] Figure 4 A flowchart illustrating another explicit tag preference update method of this application is shown.

[0061] Figure 5 A schematic diagram of the intelligent video recommendation system of this application is shown.

[0062] Figure 6 A schematic diagram of the composition structure of an electronic device according to an embodiment of this application is shown. Detailed Implementation

[0063] To make the objectives, features, and advantages of this application more apparent and understandable, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0064] Example 1:

[0065] Please refer to Figure 1 This embodiment provides an intelligent video recommendation method, including the following steps:

[0066] S1 is used to annotate the video with multiple tag groups; please refer to [link / reference]. Figure 2 The multiple tag group includes explicit tags and implicit tags;

[0067] The explicit tags represent events within the video, while the implicit tags include opinions on the explicit tags, representing the video publisher's biased opinion on the events within the video. In this embodiment, the explicit and implicit tags are represented by characters... and The markers are used to express the correspondence between the two.

[0068] The explicit labels are visible to the user, while the implicit labels are not.

[0069] Generally speaking, an event typically involves at least three different viewpoints: a positive viewpoint, a negative viewpoint, and a neutral viewpoint. If an event involves more parties, the number of viewpoints will increase.

[0070] In other words, each explicit tag corresponds to several implicit tags, but it should be noted that for a single video, its explicit tag corresponds to only one implicit tag.

[0071] Explicit tags represent specific events within the video; for example, an entertainment video might reveal the latest activities of a public figure.

[0072] Regarding the latest developments of this public figure, some video publishers will express their own biased opinions, most typically supporting or opposing the latest developments of the public figure, or expressing neutrality; it is worth mentioning that if the video publisher does not express their own biased opinion, they are also considered neutral.

[0073] The explicit and implicit tags can be manually labeled, or they can be achieved by analyzing the text, images, and other multimodal information of the video using artificial intelligence and big data.

[0074] Text-based analysis: Tags can be generated by analyzing textual information such as video titles and descriptions. For example, a CRF model can be used to extract candidate tags, and then an attention-based semantic similarity model can be used to rank the candidate tags and select appropriate ones for labeling. Alternatively, a combination of generative and extractive methods based on Transformers can be used. Generative models first generate tags, and if no suitable tags are found, extractive results are used, thereby improving the accuracy of tag labeling.

[0075] Image recognition-based methods: For cover images or video frames, pre-trained image models such as Xception can be used for feature extraction. High-frequency abstract labels are selected as image classification labels, an image classification model is constructed and fine-tuned, and then intermediate layer vectors are extracted as image representations. Relevant labels are generated based on image features.

[0076] It is worth noting that implicit labels, which represent biased opinions about events in the video, are often presented in the form of text or voice. Therefore, it is preferable to use a method based directly on text analysis or to first convert the voice into text information and then use text analysis to label the implicit labels.

[0077] S2, construct an explicit tag preference library for the user, the explicit tag preference library stores the user's preference for each explicit tag.

[0078] When building the explicit tag preference library, each explicit tag is assigned an initial preference score. It should be noted that each time a new explicit tag is generated, the new explicit tag is updated to the explicit tag preference library, and an initial preference score is assigned at the same time.

[0079] S3, responding to the user's historical video browsing history, updates the user's preference for each explicit tag stored in the explicit tag preference library.

[0080] For details, please refer to Figure 3 This includes the following steps:

[0081] S301, Set a first time threshold and a second time threshold, wherein the first time threshold is less than the second time threshold;

[0082] S302, Extract the user's historical video browsing records within a specified time period; the specified time period is preferably the most recent 10 days or 30 days.

[0083] S303, Statistically analyze the browsing time and explicit tags in the multiple tag groups of each video in the historical video browsing records;

[0084] S304, if the video viewing time is greater than the first time threshold, then increase the user's preference for those explicit tags in the explicit tag preference library; if the video viewing time is less than the first time threshold, then decrease the user's preference for those explicit tags in the explicit tag preference library; if the video viewing time is between the first time threshold and the second time threshold, then keep the user's preference for those explicit tags unchanged in the explicit tag preference library.

[0085] The time a user spends watching a video largely reflects whether they like it; the longer the viewing time, the more the user likes the video.

[0086] In this embodiment, the user's preference for explicit tags needs to be updated for all videos viewed by the user within the specified time period, following the steps in S304 above. That is, if all five videos viewed by the user within the specified time period contain the same explicit tag, then the user's preference for that explicit tag needs to be reduced five times within the explicit tag preference library.

[0087] With explicit labels For example, its initial preference level during the specified period was Each time the user's preference for these explicit tags is increased, it is... Each time, the user's preference for these explicit tags is reduced by ; and They can be equal or unequal.

[0088] Within the explicit tag preference library, users' preferences for these explicit tags... Preference Updated to:

[0089]

[0090] in, This represents the number of times a user's liking for the explicit tag has increased. This represents the number of times a user's preference for that explicit tag has been reduced.

[0091] In other embodiments, the method for updating the user's preference for each explicit tag stored in the explicit tag preference library may employ the following scheme.

[0092] The method for updating the user's preference level for each explicit tag in the explicit tag preference library based on the user's historical video browsing history is as follows, please refer to... Figure 4 This includes the following steps:

[0093] S310, each video is configured with a feedback mechanism, which is triggered during or after video playback to display explicit tags of the video to the user for positive or negative feedback.

[0094] This section only provides explicit tags for user feedback, which explains why explicit tags are visible to users while implicit tags are not. This is to prevent users from intentionally or unintentionally providing feedback on implicit tags, which could lead to subsequent videos only pushing a specific viewpoint on the event, thus creating an information cocoon.

[0095] The feedback mechanism is triggered in several ways, including but not limited to: long-touching the video playback interface during video playback, or automatically after video playback ends; positive / negative feedback includes liking / disliking, following / not following, etc.

[0096] S311, Set a first time threshold and a second time threshold, wherein the first time threshold is less than the second time threshold;

[0097] S312, Extract the user's historical video browsing records within a specified time period;

[0098] S313, Statistically analyze the browsing time of each video in the historical video browsing records, the explicit tags in the video's multiple tag groups, and the user's feedback on the explicit tags.

[0099] S314, If the video viewing time exceeds the first time threshold, then within the explicit tag preference library, increase the user's first preference for those explicit tags. If the video viewing time is less than the first time threshold, then within the explicit tag preference library, the user's second preference for those explicit tags will be reduced. If the video viewing time is between the first and second time thresholds, then the user's preference for those explicit tags will remain unchanged in the explicit tag preference library.

[0100] S315, If the user's feedback on the explicit labels is positive, then increase the user's third preference for those explicit labels. Conversely, negative feedback from users regarding explicit tags reduces their fourth level of liking for those tags. If the user does not respond to the explicit tag, the user's preference for that explicit tag will remain unchanged.

[0101] Similarly, for all videos viewed by the user within the specified time period, the user's preference for explicit tags needs to be updated according to steps S314 and S315 above. In S315, if the user provides feedback on the same explicit tag multiple times, the preference for the explicit tag needs to be lowered or increased accordingly multiple times.

[0102] Or use explicit labels For example, its initial preference level during the specified period was After steps S310 to S315, in the explicit tag preference library, the user selects these explicit tags... Preference Updated to:

[0103]

[0104] in, This represents the number of times a user's liking for the explicit tag has increased. This represents the number of times a user's preference for that explicit tag has been reduced; This represents the number of times a user provides positive feedback on the explicit label; This represents the number of times a user provides negative feedback on the explicit label.

[0105] It should be noted that the first preference level Second preference Third preference Fourth preference The values ​​can be different, but the following condition must be met: Third preference Greater than the first preference And the fourth preference Greater than the second preference The reason lies in the third preference. Fourth preference It's a reward / penalty system based on user-generated video feedback, and the first preference level... Second preference The reward / penalty is determined by analyzing the user's video viewing time. The former should be more reliable than the latter, so they should be distinguished in terms of value selection.

[0106] S4. The priority value of the video to be recommended is obtained by combining the current popularity of the video to be recommended with the product of the user's preference for the explicit tags of the video to be recommended and the weight factor. The weight factor is negatively correlated with the number of times the implicit tags of the video to be recommended have been viewed in the historical video browsing records.

[0107] Specifically, the following methods are included:

[0108] S401, construct weight factors that are negatively correlated with the implicit tags of the video to be recommended;

[0109] Extract the user's historical video browsing history within a specified time period and analyze the implicit tags containing videos to be recommended. The corresponding explicit tag The video in question is designated as the first video.

[0110] The total number of first videos and the implicit tags containing the video to be recommended. The number of first videos ;

[0111] Calculate the number of first videos containing the implicit tags of the video to be recommended. The total number of videos with the first video The ratio is calculated by subtracting the ratio from the constant to obtain the weighting factor. The constant is usually taken as 1, that is:

[0112]

[0113] S402, calculate the current popularity of the video to be recommended;

[0114] Extract explicit tags from the video to be recommended, and count the number of times all users have viewed the video within a specified time period, recording this as the first time count. ;

[0115] It should be noted that the criteria for determining whether a user has viewed the video to be recommended are as follows:

[0116] If a user spends more than a first time threshold browsing the video to be recommended, and / or if the user provides positive feedback on the video to be recommended, then the user is considered to have viewed the video to be recommended.

[0117] With the first number Divide by the set number of times threshold This involves normalizing the data to obtain the videos to be recommended. Current popularity .

[0118] S403, Calculate the current popularity of the video to be recommended. And users' preferences for explicit tags in recommended videos. With weighting factors The sum of the products is used as the priority value for obtaining the video to be recommended. .

[0119] .

[0120] S5, according to priority value The videos to be recommended are sorted by priority from highest to lowest, and the videos with the highest priority are selected for recommendation.

[0121] This means that before the first batch of recommended videos finishes playing, the next round of recommended videos is prioritized and recommended to ensure a smooth viewing experience for users.

[0122] Through the above steps, due to the weighting factor The implicit tags of a video to be recommended have a negative correlation with the number of times they are viewed in the historical video browsing history. Therefore, if the implicit tags of a video to be recommended account for a higher percentage of the historical video browsing history, the priority value of videos to be recommended with the same implicit tags will be reduced in subsequent recommendations. Conversely, if the implicit tags of a video to be recommended account for a lower percentage of the historical video browsing history, the priority value of videos to be recommended with the same implicit tags will be increased in subsequent recommendations.

[0123] This solution can push relevant videos to users based on video popularity and the preference of explicit tags. It can also combine a weighting factor that is negatively correlated with the number of times the implicit tags of the recommended videos are viewed in the historical video browsing records. This makes it easier for users to see videos with different implicit tags when they browse videos with the same explicit tags later. In other words, for the same event, it can obtain different public opinions on it, avoiding the limitation of the understanding of the same event to a single viewpoint. This is especially helpful for children, the elderly and other groups, making it less likely for them to fall into the predicament of information cocoon.

[0124] While various video uploaders may present different perspectives on the same event, a significant portion of their videos lack detailed information and contain biased viewpoints. This makes it difficult for viewers to gain a complete understanding of the event, and they may even be misled by the opinions of some video uploaders.

[0125] To avoid this situation, the solution also includes the following methods:

[0126] If a user spends more than a first time threshold browsing a video marked with a certain explicit tag more than a set number of times, and / or, the user provides positive feedback on a certain explicit tag more than a set number of times, the user will be marked as a user followed by that explicit tag.

[0127] In S1, appropriate explicit and notification tags are added to event-related notification videos published by certified video publishers;

[0128] Certified video publishers are usually government agencies at various levels, such as public security bureaus, courts, and other administrative organs.

[0129] In S5, when prioritizing videos, if one of the videos to be recommended is marked with a notification tag, the video marked with the notification tag will be given the highest priority and ranked first when recommended to users who follow the corresponding explicit tag.

[0130] Example 2:

[0131] This second embodiment provides an intelligent video recommendation system. Please refer to... Figure 5 It includes the following functional modules:

[0132] The tagging module is used to annotate videos with multiple tag groups; the multiple tag groups include explicit tags and implicit tags;

[0133] The preference building module builds an explicit tag preference library for users, which stores the user's preference for each explicit tag.

[0134] The preference update module responds to the user's historical video browsing history and updates the user's preference for each explicit tag stored in the explicit tag preference library;

[0135] The priority calculation module combines the current popularity of the video to be recommended with the product of the user's preference for the explicit tags of the video to be recommended and the weight factor to obtain the priority value of the video to be recommended. The weight factor is negatively correlated with the number of times the implicit tags of the video to be recommended have been viewed in the historical video browsing records.

[0136] The sorting and recommendation module sorts the videos to be recommended according to their priority values ​​from highest to lowest, and selects the first number of videos with the highest priority to recommend.

[0137] The specific implementation methods / principles of each functional module in this embodiment are the same as those in embodiment one. For details, please refer to steps S1 to S5 in embodiment one, which will not be repeated here.

[0138] Example 3

[0139] This third embodiment also provides an electronic device and a readable storage medium.

[0140] Figure 6 A schematic block diagram of an example electronic device that can be used to implement embodiments of this application is shown. The electronic device 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 electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the application described and / or claimed herein.

[0141] like Figure 6 As shown, the device includes a computing unit that can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) or a computer program loaded from a storage unit into random access memory (RAM). The RAM can also store various programs and data required for device operation. The computing unit, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0142] Multiple components in the device are connected to the I / O interface, including: input units such as keyboards and mice; output units such as various types of displays and speakers; storage units such as disks and optical discs; and communication units such as network interface cards (NICs), modems, and wireless transceivers. The communication unit allows the device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0143] The computing unit can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of computing units include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit performs the various methods and processes described above, such as the intelligent video recommendation method described in Embodiment 1. For example, in some embodiments, the intelligent video recommendation method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as a storage unit. In some embodiments, part or all of the computer program may be loaded and / or installed on the device via ROM and / or a communication unit. When the computer program is loaded into RAM and executed by the computing unit, one or more steps of the intelligent video recommendation method described above may be performed. Alternatively, in other embodiments, the computing unit may be configured to perform the intelligent video recommendation method by any other suitable means (e.g., by means of firmware).

[0144] Various embodiments 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-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0145] The program code used to implement the methods of this application may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0146] In the context of this application, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0147] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0148] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0149] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0150] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this application can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this application can be achieved, and this is not limited herein.

[0151] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "a plurality of" means two or more, unless otherwise explicitly specified.

[0152] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An intelligent video recommendation method, characterized in that, Includes the following steps: S1, labeling the video with multiple tag groups; the multiple tag groups include explicit tags and implicit tags; the explicit tags represent events within the video, and the implicit tags include opinions on the explicit tags; S2, Build an explicit tag preference library for the user, wherein the explicit tag preference library stores the user's preference for each explicit tag; S3, responding to the user's historical video browsing history, updates the user's preference for each explicit tag stored in the explicit tag preference library; S4. The priority value of the video to be recommended is obtained by combining the current popularity of the video to be recommended with the product of the user's preference for the explicit tags of the video to be recommended and the weight factor. The weight factor is negatively correlated with the number of times the implicit tags of the video to be recommended have been viewed in the historical video browsing records. S5: Sort the videos to be recommended according to their priority values ​​from highest to lowest, and select the videos with the highest priority values ​​to recommend. In step S3, the method for updating the user's preference for each explicit tag stored in the explicit tag preference library in response to the user's historical video browsing history is as follows: S301, Set a first time threshold and a second time threshold, wherein the first time threshold is less than the second time threshold; S302, Extract the user's historical video browsing records within a specified time period; S303, Statistically analyze the browsing time and explicit tags in the multiple tag groups of each video in the historical video browsing records; S304, if the video viewing time is greater than the first time threshold, then increase the user's preference for those explicit tags in the explicit tag preference library; if the video viewing time is less than the first time threshold, then decrease the user's preference for those explicit tags in the explicit tag preference library; if the video viewing time is between the first time threshold and the second time threshold, then keep the user's preference for those explicit tags unchanged in the explicit tag preference library. In step S4, the method for obtaining the priority value of the video to be recommended by combining the current popularity of the video to be recommended with the product of the user's preference for the explicit tags of the video to be recommended and the weight factor is as follows: S401, construct weight factors that are negatively correlated with the implicit tags of the video to be recommended; S402, calculate the current popularity of the video to be recommended; S403, calculate the current popularity of the video to be recommended, and the sum of the product of the user's preference for the explicit tags of the video to be recommended and the weight factor, as the priority value for obtaining the video to be recommended.

2. The intelligent video recommendation method according to claim 1, characterized in that, In step S2, when constructing the explicit tag preference library, each explicit tag is assigned an initial preference level; and each time a new explicit tag is generated, the new explicit tag is updated to the explicit tag preference library, and an initial preference level is assigned at the same time.

3. The intelligent video recommendation method according to claim 1, characterized in that, In step S3, the method for updating the user's preference for each explicit tag stored in the explicit tag preference library in response to the user's historical video browsing history is as follows: S310, each video is configured with a feedback mechanism, which is triggered during or after video playback to display explicit tags of the video to the user for positive or negative feedback. S311, Set a first time threshold and a second time threshold, wherein the first time threshold is less than the second time threshold; S312, Extract the user's historical video browsing records within a specified time period; S313, Statistically analyze the browsing time of each video in the historical video browsing records, the explicit tags in the video's multiple tag groups, and the user's feedback on the explicit tags. S314, if the video viewing time is greater than the first time threshold, then within the explicit tag preference library, increase the user's first preference for these explicit tags; if the video viewing time is less than the first time threshold, then within the explicit tag preference library, decrease the user's second preference for these explicit tags; if the video viewing time is between the first time threshold and the second time threshold, then within the explicit tag preference library, keep the user's preference for these explicit tags unchanged. S315, if the user provides positive feedback on the explicit labels, the user's third preference for those explicit labels increases; conversely, if the user provides negative feedback on the explicit labels, the user's fourth preference for those explicit labels decreases; if the user does not provide feedback on the explicit labels, the user's preference for those explicit labels remains unchanged.

4. The intelligent video recommendation method according to claim 1, characterized in that, In step S401, the method for constructing weight factors that are negatively correlated with the implicit tags of the video to be recommended is as follows: Extract the user's historical video browsing history within a specified time period, and count the videos that contain the explicit tags corresponding to the implicit tags of the videos to be recommended, and record them as the first video; The total number of first videos and the number of first videos containing the implicit tags of the video to be recommended are counted. Calculate the ratio of the number of first videos containing the implicit tag of the video to be recommended to the total number of first videos, and subtract this ratio as a constant to obtain the weighting factor.

5. The intelligent video recommendation method according to claim 1, characterized in that, In step S402, the method for calculating the current popularity of the video to be recommended is as follows: Extract the explicit tags of the video to be recommended, and count the number of times all users have viewed the video to be recommended within the specified time period, which is recorded as the first time count; Divide the first number of times by the set number threshold to obtain the current popularity of the video to be recommended.

6. The intelligent video recommendation method according to claim 5, characterized in that, The criteria for determining whether a user has viewed the video to be recommended are as follows: when the user's viewing time of the video to be recommended exceeds a first time threshold, and / or the user provides positive feedback on the video to be recommended, then the user is considered to have viewed the video to be recommended.

7. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.

8. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-6.

Images