A short video playback recommendation method and device based on big data

Abstract

The present application relates to the technical field of video recommendation, and in particular to a short video playback recommendation method and device based on big data, which collects metadata of a video to be recommended; obtains corresponding POI data of each candidate interest segment in M ​​candidate interest segments in the video to be recommended through metadata; obtains user history data of each candidate interest segment in the M candidate interest segments through POI data of each candidate interest segment in the M candidate interest segments; obtains corresponding interest segments through user history data of each candidate interest segment in the M candidate interest segments and corresponding segment location requests of each candidate interest segment in the M candidate interest segments, so as to obtain M candidate interest segments; locates and intercepts the M candidate interest segments, so as to obtain the video to be recommended. The present invention can effectively locate the interest of the segments of the video to be recommended based on the user's viewing history, and provide suitable recommended segments.

Description

Technical Field

[0001] The present application relates to the technical field of video recommendation, and in particular to a short video playback recommendation method and device based on big data. Background Art

[0002] There can be many specific services in short videos. For example, there can be a service for pushing recommended content. The server obtains the user's historical click records, determines the type of content that the user likes based on the historical click records, searches for content related to the type, and pushes the related content as recommended content to the terminal display interface for the user to browse.

[0003] At present, when collecting users' discussions on short videos that have been broadcast, staff usually quickly browse the comments on major websites, filter out the hotly discussed content, and then summarize it to obtain users' feedback on the short videos that have been broadcast. However, this manual collection method is inefficient and takes a long time, and it may not be possible to obtain useful information in a short time; on the other hand, some users may make random comments without watching the short video, resulting in the collected feedback being inaccurate.

[0004] Therefore, according to the current method of collecting user information, users may not be able to quickly locate the segments of interest when watching short videos, resulting in low viewing efficiency. Summary of the invention

[0005] To achieve the above objectives, this application provides the following technical solutions:

[0006] According to a first aspect of the present invention, the present invention seeks protection for a short video playback recommendation method based on big data, the method comprising:

[0007] Collect metadata of the video to be recommended;

[0008] Obtaining POI data corresponding to each of the M candidate interest segments in the video to be recommended through the metadata;

[0009] Obtaining user history record data of each candidate interest segment in the M candidate interest segments through the POI data of each candidate interest segment in the M candidate interest segments;

[0010] Obtain a corresponding interest segment through user history record data of each candidate interest segment in the M segments and a segment location request corresponding to each candidate interest segment in the M segments, so as to obtain a target interest segment;

[0011] The target interest segment is located and intercepted to obtain the video to be recommended.

[0012] Further, the step of obtaining a corresponding interest segment through user history record data of each candidate interest segment in the M candidate interest segments and a corresponding segment location request to obtain a target interest segment includes:

[0013] Collecting historical viewing records corresponding to each candidate interest segment in the M candidate interest segments through user history record data of each candidate interest segment in the M candidate interest segments;

[0014] Performing stop viewing time classification processing on the M candidate interest segments to obtain P stop viewing POI lists, wherein each stop viewing POI list in the P stop viewing POI lists has a different stop viewing time, and the candidate interest segments in the stop viewing POI lists have the same stop viewing time;

[0015] According to the corresponding response order of the stop viewing time of the P stop viewing POI lists, the corresponding historical viewing records of the candidate interest segments in the stop viewing POI lists are run in the corresponding element library of the P stop viewing POI lists and the corresponding segment positioning request is responded to to obtain the target interest segment.

[0016] Further, before executing the corresponding historical viewing records of the candidate interest segments in the stop viewing POI list in the element library corresponding to the P stop viewing POI list according to the response order corresponding to the stop viewing duration of the P stop viewing POI list and responding to the corresponding segment positioning request to obtain the target interest segment, the method further includes:

[0017] Collecting video access data of historical viewing records corresponding to each candidate interest segment in the M candidate interest segments;

[0018] If the video access frequency indicated by the video access data of the historical viewing record corresponding to the first candidate interest segment among the M candidate interest segments is lower than the video access frequency threshold indicated by the preset video access data, sending history request data to the history server, the history request data including the preset video access data corresponding to the first candidate interest segment;

[0019] Receiving historical viewing records corresponding to preset video access data sent by the history server;

[0020] The historical viewing record corresponding to the preset video access data is obtained as the historical viewing record corresponding to the first candidate interest segment.

[0021] Furthermore, the method further comprises:

[0022] Collecting request response results corresponding to each candidate interest segment in the M candidate interest segments;

[0023] Obtaining corresponding response status data of each candidate interest segment in the M segments of candidate interest segments through the corresponding request response result of each candidate interest segment in the M segments of candidate interest segments;

[0024] Obtaining a second candidate interest segment, where the second candidate interest segment is a candidate interest segment whose response status indicated by the response status data is a priority status among the M candidate interest segments;

[0025] collecting a second response parameter of the second candidate interest segment;

[0026] Determining a priority response factor of the second candidate interest segment by using the second response parameter;

[0027] Determining the priority level of the second candidate interest segment according to the priority response factor;

[0028] The priority is displayed.

[0029] Further, after deriving the priority response factor of the second candidate interest segment through the second response parameter, the method further includes:

[0030] If the priority factor indicated by the priority response factor is video plot priority, then collecting current video plot status data;

[0031] If the current video plot status data is the priority video plot status data, then displaying the current video plot status data;

[0032] Receiving video plot data corresponding to the second candidate interest segment input by a candidate user;

[0033] The second candidate interest segment is obtained through the historical viewing record corresponding to the second candidate interest segment, the segment location request and the video plot data.

[0034] Further, after deriving the priority response factor of the second candidate interest segment through the second response parameter, the method further includes:

[0035] If the priority factor indicated by the priority response factor is video character priority, collecting a first priority video character from the first candidate video characters of the second candidate interest segment by using the second response parameter;

[0036] Performing a priority analysis on the first priority video character to obtain a priority analysis result;

[0037] Adjusting the first priority video character according to the priority analysis result to obtain a second video character;

[0038] Replacing the first priority video character in the first candidate video character corresponding to the second candidate interest segment with the second video character to obtain a second candidate video character;

[0039] The second candidate interest segment is obtained through the second candidate video character, the historical viewing record corresponding to the second candidate interest segment, and the segment positioning request.

[0040] Furthermore, the method further comprises:

[0041] Determine a first priority through the priority analysis result;

[0042] If the first priority is higher than the preset priority, the first priority video character and the second priority video character are displayed.

[0043] According to a second aspect of the present invention, the present invention seeks protection for a short video playback recommendation device based on big data, the device comprising:

[0044] A collection module, used to collect metadata of the video to be recommended;

[0045] A first output module is used to obtain POI data corresponding to each of the M candidate interest segments in the video to be recommended through the metadata;

[0046] A second output module is used to obtain user history record data of each candidate interest segment in the M candidate interest segments through POI data of each candidate interest segment in the M candidate interest segments;

[0047] A third output module is used to obtain a corresponding interest segment through user history record data of each candidate interest segment in the M segments and a segment location request corresponding to each candidate interest segment in the M segments, so as to obtain a target interest segment;

[0048] A positioning and interception module, used for positioning and intercepting the target interest segment to obtain the video to be recommended;

[0049] The short video playback recommendation device based on big data is used to execute the short video playback recommendation method based on big data.

[0050] The present application relates to the technical field of video recommendation, and in particular to a short video playback recommendation method and device based on big data, which collects metadata of a video to be recommended; obtains corresponding POI data of each candidate interest segment in M ​​candidate interest segments in the video to be recommended through metadata; obtains user history data of each candidate interest segment in the M candidate interest segments through POI data of each candidate interest segment in the M candidate interest segments; obtains corresponding interest segments through user history data of each candidate interest segment in the M candidate interest segments and corresponding segment location requests of each candidate interest segment in the M candidate interest segments, so as to obtain M candidate interest segments; locates and intercepts the M candidate interest segments, so as to obtain the video to be recommended. The present invention can effectively locate the interest of the segments of the video to be recommended based on the user's viewing history, and provide suitable recommended segments. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A schematic diagram of the structure of a short video playback recommendation device based on big data as claimed in an embodiment of the present application;

[0052] Figure 2 A workflow diagram of a short video playback recommendation method based on big data as claimed in an embodiment of the present application;

[0053] Figure 3 This is a structural module diagram of a short video playback recommendation device based on big data that is requested to be protected by an embodiment of the present application. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0055] The terms "first", "second" and "third" in this application are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined as "first", "second" and "third" can explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. All directional indications (such as up, down, left, right, front, back...) in the embodiments of this application are only used to explain the relative position relationship, movement, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication also changes accordingly. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, device, product or equipment that includes a series of steps or modules is not limited to the listed steps or modules, but optionally also includes steps or modules that are not listed, or optionally also includes other steps or modules inherent to these processes, methods, products or equipment.

[0056] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0057] See also Figure 1 , Figure 1 The present application provides a schematic diagram of a short video playback recommendation device based on big data. Figure 1As shown, the device includes a server and a client. The user can input metadata of the video to be recommended on the client. The metadata can be configuration data of the video to be recommended, etc. The metadata at least includes feature data of each video frame of the video to be recommended, such as semantic features, color features, label features, content features, etc. The client sends the metadata to the server. The server obtains the POI data corresponding to each interest segment through the metadata, obtains the user history data of each interest segment through the POI data, and locates the corresponding interest segment through the user history data and the corresponding segment location request, and finally locates and intercepts the interest segment to obtain the video to be recommended. Among them, the application segment location request can be a request for creating an application segment. After responding to the request, the corresponding source code can be run to locate the segment, so as to obtain the corresponding segment. Therefore, there is no need to manually collect the viewing history corresponding to the candidate interest segment, and the candidate interest segment can be quickly recommended, which improves the efficiency of locating the candidate interest segment, and further improves the efficiency of locating the video.

[0058] See also Figure 2 As shown, Figure 2 A flowchart of a short video playback recommendation method based on big data provided by an embodiment of the present invention. Figure 2 As shown, the short video playback recommendation method based on big data includes the following steps:

[0059] S201: Collect metadata of the video to be recommended.

[0060] The metadata of the video to be recommended can be collected from the configuration data of the video to be recommended. The metadata may include the POI data package name, video access number, etc. The video to be recommended may be a short video, such as a micro-movie, a TV mini-series, etc. The configuration data of the video to be recommended may be pre-configured by the user when locating the candidate application, including indication data indicating the metadata of the video to be recommended, so that the corresponding metadata can be collected through the corresponding indication data.

[0061] S202: Obtain POI data corresponding to each of the M candidate interest segments in the video to be recommended through the metadata.

[0062] Among them, the POI data corresponding to each interest segment can be obtained through the POI data packet name and the corresponding video access number in the POI metadata. The POI data packet name can be the POI data packet name corresponding to the K candidate interest segments respectively. The POI data packet name and the corresponding video access number can uniquely obtain the POI data corresponding to the candidate interest segment, and the POI data can include data describing the script of the candidate interest segment (script description data), and identification data of the candidate interest segment, etc. The script description data may include indication data indicating the third-party database, third-party interface, etc. that it is watching, as well as video access data of the third-party database watched, etc.

[0063] S203. Obtain user history record data of each candidate interest segment in the M candidate interest segments through the POI data of each candidate interest segment in the M candidate interest segments.

[0064] The user history record data can be obtained through the script description data in the POI data of each candidate interest segment. Specifically, the user history record data can be collected through the indication data of the third-party database, third-party data interface, etc. indicating the viewing of the candidate interest segment in the script description data, and the video access data of the viewed third-party database, etc.

[0065] S204, deriving a corresponding interest segment through the user history record data of each candidate interest segment in the M segments and the corresponding segment location request of each candidate interest segment in the M segments, so as to obtain a target interest segment.

[0066] Among them, after collecting the user history record data, the video access data of the third-party database, the identification data of the third-party database, etc. in the user history record data can be collected into the corresponding third-party database related data in the server cache, for example, the historical package of the third-party database, etc.

[0067] After collecting historical packages from a third-party database, etc., respond to the segment location request in the thread corresponding to the candidate interest segment to locate the segment. When locating the segment, respond to the video characters corresponding to the candidate interest segment and call the history in the historical package of the third-party database to locate the segment to obtain the candidate interest segment.

[0068] S205: Position and intercept the target interest segment to obtain the video to be recommended.

[0069] The logical data of the M candidate interest segments may be used to locate and intercept the video to be recommended. The logical data of the M candidate interest segments may be understood as locating and intercepting the M candidate interest segments in the video to be recommended, thereby obtaining the video to be recommended.

[0070] Specifically, the positioning interception can be distinguished by identifying features in the video to be recommended, for example, by using the scene transition as the positioning interception point or the change in the number of characters as the positioning interception point.

[0071] In this example, metadata of the video to be recommended is collected, and the corresponding POI data of each candidate interest segment in the M candidate interest segments in the video to be recommended is obtained through the metadata, and the user history record data of each candidate interest segment in the M candidate interest segments is obtained through the POI data of each candidate interest segment in the M candidate interest segments, and the corresponding interest segment is obtained through the user history record data of each candidate interest segment in the M candidate interest segments and the corresponding segment positioning request to obtain the target interest segment, and the target interest segment is located and intercepted to obtain the video to be recommended. In the present invention, in the positioning of video data for the short video playback platform, the corresponding candidate interest segment can be obtained by collecting the corresponding user history record data of each candidate interest segment in the video to be recommended and the corresponding segment positioning request, and finally the video to be recommended can be obtained by locating and intercepting multiple candidate interest segments, so that the candidate interest segment can be quickly recommended without manually collecting the viewing history corresponding to the candidate interest segment, thereby improving the efficiency of locating the candidate interest segment, and further improving the efficiency of locating the video.

[0072] In a possible implementation, in the existing video positioning method, when positioning each interest segment, a serial response method is usually used for positioning, so that the currently unpositioned interest segment needs to wait until the positioning of the interest segment being positioned is completed before positioning, resulting in low efficiency in positioning the interest segment. In order to solve the above problem, the embodiment of the present application provides a method for quickly positioning the interest segment to improve the efficiency of positioning the interest segment, which can be specifically: a possible method of obtaining the corresponding interest segment through the user history record data of each candidate interest segment in the M segments of the candidate interest segment and the corresponding segment positioning request to obtain the corresponding interest segment, including:

[0073] A1. Collecting historical viewing records corresponding to each candidate interest segment in the M candidate interest segments through user history record data of each candidate interest segment in the M candidate interest segments;

[0074] A2, classifying the viewing duration of the M candidate interest segments to obtain P viewing POI lists, wherein each of the P viewing POI lists has a different viewing duration, and the viewing duration of the candidate interest segments in the viewing POI lists is the same;

[0075] A3. According to the corresponding response order of the viewing duration of the P viewing POI lists, run the corresponding historical viewing records of the candidate interest segments in the viewing POI lists in the element libraries corresponding to the P viewing POI lists and respond to the corresponding segment positioning requests to obtain the target interest segment.

[0076] Among them, the relevant data of the corresponding third-party database, such as the history package of the third-party database, can be collected in the cache of the server through the video access data of the third-party database and the identification data of the third-party database in the user history record data. Of course, the corresponding data can also be collected from the corresponding third-party data interface through the third-party data interface data in the user history record data to obtain the user history record data. Of course, the above two methods can also respond at the same time, specifically, whether the user history record data includes the video access data of the third-party database, the third-party database interface data, etc. If included, they need to be collected at the same time.

[0077] Since the operation history of the server for locating the candidate interest segments is limited, it may be impossible to locate each of the M candidate interest segments in parallel at the same time. Therefore, positioning can be performed in batches. Specifically, the viewing time of each candidate interest segment can be classified and processed to obtain P viewing POI lists. The viewing time of the candidate interest segments in each viewing POI list is the same, but the viewing time between the viewing POI lists is different. Therefore, positioning can be performed in batches to improve the efficiency of positioning.

[0078] In a specific example, there are 9 candidate interest segments, which are divided into 3 stay viewing POI lists according to their corresponding stay viewing time, namely, the first stay viewing POI list, the second stay viewing POI list, and the third stay viewing POI list, which correspond to the first stay viewing time, the second stay viewing time, and the third stay viewing time, respectively. The response priority of the first stay viewing time is greater than the second stay viewing time, and the response priority of the second stay viewing time is greater than the third stay viewing time. At this time, the segment location request of the candidate interest segment in the first stay viewing POI list can be responded to in the corresponding thread first, and the corresponding third-party database history package can be run to obtain the corresponding candidate interest segment. After the candidate interest segment in the first stay viewing POI list is located, the candidate interest segment in the second stay viewing POI list is located, and finally the candidate interest segment in the third stay viewing POI list is located. Therefore, compared with the time consumed when the 9 candidate interest segments are responded to serially, the parallel response method can greatly reduce the candidate interest segment location time, thereby improving the location efficiency of the video to be recommended.

[0079] In a possible implementation, before locating the candidate interest segment, it is also necessary to perform a video access check on the historical viewing records. If there is a video access of the historical viewing records that cannot meet the requirements, it is necessary to update it to improve the reliability of the corresponding candidate interest segment positioning. Specifically, it can be: according to the corresponding response order of the viewing duration of the P viewing POI lists, the corresponding historical viewing records of the candidate interest segments in the viewing POI list are run in the element library corresponding to the P viewing POI lists and the corresponding segment positioning request is responded to, so as to obtain the target interest segment, and then the video access update check is performed on the historical viewing records, which is as follows:

[0080] B1, collecting video access data of historical viewing records corresponding to each candidate interest segment in the M candidate interest segments;

[0081] B2. If the video access frequency indicated by the video access data of the historical viewing record corresponding to the first candidate interest segment among the M candidate interest segments is lower than the video access frequency threshold indicated by the preset video access data, send historical request data to the history server, wherein the historical request data includes the preset video access data corresponding to the first candidate interest segment;

[0082] B3, receiving the historical viewing record corresponding to the preset video access data sent by the history server;

[0083] B4. Derive the historical viewing record corresponding to the preset video access data as the historical viewing record corresponding to the first candidate interest segment.

[0084] The preset video access data may be the latest video access of the historical viewing record, which may be specifically indicated by a video access number. Since the historical viewing record stored in the cache of the server may not be the latest video access, the server may collect video access data corresponding to the historical viewing record at a certain time interval. If the video access of the historical viewing record is not the latest video access when locating the candidate interest segment, the historical viewing record corresponding to the latest video access needs to be collected.

[0085] In this example, if there are historical viewing record video accesses that cannot meet the demand, they need to be updated to improve the reliability of locating the corresponding candidate interest segments.

[0086] In a possible implementation, the request response result of the candidate interest segment may also be analyzed and processed, and a reminder process may be performed when the response status of the candidate interest segment is a priority status, as follows:

[0087] C1. Collect the request response results corresponding to each candidate interest segment in the M candidate interest segments;

[0088] C2. Obtaining corresponding response status data of each candidate interest segment in the M candidate interest segments through the corresponding request response result of each candidate interest segment in the M candidate interest segments;

[0089] C3. Obtain a second candidate interest segment, where the second candidate interest segment is a candidate interest segment whose response status indicated by the response status data is a priority status among the M candidate interest segments;

[0090] C4. Collecting a second response parameter of the second candidate interest segment;

[0091] C5. deriving a priority response factor of the second candidate interest segment according to the second response parameter;

[0092] C6. deriving a priority level of the second candidate interest segment according to the priority response factor;

[0093] C7. Display the priority.

[0094] Among them, the request response results of the candidate interest fragments can be collected from the cache. The request response result can represent the request response status, and different request response structures can indicate different request response statuses. For example, if the request response result is the same as the preset request response result, the request response status data can be a normal response; if the request response result is different from the preset request response result, the request response status data can be a response priority, etc.

[0095] The second response parameters of the second candidate interest segment may include video plot parameters, operating environment parameters, operating parameters of historical viewing records, response logs of video characters of the second candidate POI, etc. when responding.

[0096] The second response parameter may be analyzed with priority to obtain a priority response factor. Specifically, for example, the video plot parameter may be analyzed to determine whether it is a normal video plot parameter. If the video plot parameter is prioritized, the priority response factor may be a response factor that characterizes the priority of the video plot. For another example, the response log of the video character of the second candidate POI may be analyzed to determine whether the video character of the second candidate POI is prioritized. If the video character is prioritized through the log, the priority response factor may be a response factor that characterizes the priority of the video character.

[0097] Different priority response factors have their corresponding priorities, so the priority of the second candidate interest segment can be obtained through the priority response factor and displayed. The priority can be displayed through SMS, email, pop-up window, etc. This is only an example and is not specifically limited.

[0098] In this example, the request response result of the candidate interest segment can be analyzed and processed, and a reminder process can be performed when the response status of the candidate interest segment is a priority status, thereby improving the reliability of locating the candidate interest segment.

[0099] In a possible implementation, if the priority factor indicated by the priority response factor is video plot priority, the corresponding video plot data can be collected separately and subsequently positioned, thereby improving the reliability of positioning the candidate interest segment. The method is as follows:

[0100] D1. If the priority factor indicated by the priority response factor is video plot priority, then collect current video plot status data;

[0101] D2. If the current video plot status data is the priority video plot status data, then display the current video plot status data;

[0102] D3, receiving video plot data corresponding to the second candidate interest segment input by the candidate user;

[0103] D4. Obtain the second candidate interest segment through the historical viewing record corresponding to the second candidate interest segment, the segment location request and the video plot data.

[0104] Among them, since the priority factor is video plot priority, the video plot priority indicates that the video plot is prioritized when locating the second candidate interest segment, and the current video plot state data can be collected. If the video plot state indicated by the current video plot state data is still the priority video plot state, the current video plot state data can be displayed.

[0105] After the candidate user has collected the current video plot state data as a priority, he can manually collect the video plot data corresponding to the second candidate interest segment and input it into the server. Specifically, the video plot data can be input into the server through the client corresponding to the candidate user.

[0106] After receiving the video plot data, the server obtains the second candidate interest segment through the historical viewing record corresponding to the second candidate interest segment, the segment location request and the video plot data. The specific method for obtaining the second candidate interest segment can refer to the method for obtaining the candidate interest segment in the above embodiment, which will not be repeated here.

[0107] Of course, if the video plot state indicated by the current video plot state data is a normal video plot state, the video plot data corresponding to the second candidate interest segment can be directly collected from the video plot and subsequently processed.

[0108] In this example, if the priority factor indicated by the priority response factor is video plot priority, the corresponding video plot data can be collected separately and subsequently positioned, thereby improving the reliability of positioning the candidate interest segments, and further improving the reliability of positioning the video to be recommended.

[0109] In a possible implementation, if the priority factor indicated by the priority response factor is video character priority, the corresponding priority video character can be collected and adjusted, and then the second candidate interest segment is located by the adjusted video character to obtain the second candidate interest segment, thereby improving the reliability of the second candidate interest segment location, as follows:

[0110] E1. If the priority factor indicated by the priority response factor is video character priority, collecting a first priority video character from the first candidate video characters of the second candidate interest segment by using the second response parameter;

[0111] E2. Prioritize analysis of the first priority video character to obtain a priority analysis result;

[0112] E3. Adjusting the first priority video character according to the priority analysis result to obtain a second video character;

[0113] E4. Replace the first priority video character in the first candidate video character corresponding to the second candidate interest segment with the second video character to obtain a second candidate video character;

[0114] E5. Obtain the second candidate interest segment through the second candidate video character, the historical viewing record corresponding to the second candidate interest segment, and the segment location request.

[0115] The response parameter may be a response log of a video character of the second candidate POI, and the corresponding first priority video character may be collected through the priority points in the response log.

[0116] The method for performing a priority analysis on the first priority video character may be to collect script data corresponding to the first priority video character; obtain multiple reference video characters through the script data; collect difference data between the first priority video character and the multiple reference video characters; obtain priority data of the first priority video character through the difference data to obtain a priority analysis result. The multiple reference video characters are video characters that correspond to the script data and can operate normally. Different difference data may correspond to different priority data.

[0117] The method of adjusting the first priority video character by the priority analysis result and obtaining the second video character may be: adjusting the first priority video character by the difference data corresponding to the priority analysis result. Specifically, it may be: through the difference data, the corresponding part of the first priority video character is replaced by the part in the difference data that is different from the first priority video character, and then the parameters of the replacement part are set to obtain the second video character. The method of setting the parameters of the replacement part may be: the parameters in the part of the first priority source in the difference data are correspondingly set to the parameters of the part in the difference data that is different from the first priority video character. For example, the difference data are M1 and M2, M1 is the video character in the first priority video character, and M2 is the video character in the reference video character, then M1 in the first video character is replaced by M2, and the parameters of M1 are set to the parameters of the corresponding part in M2, thereby completing the adjustment of the first priority video character and obtaining the second video character.

[0118] The method of deriving the candidate interest segment through the second candidate video character, the historical viewing record corresponding to the second candidate interest segment, and the segment positioning request can refer to the method of deriving the candidate interest segment in the aforementioned embodiment, which will not be repeated here.

[0119] In this example, if the priority response factor indicates that the priority factor is video character priority, the corresponding priority video character can be collected and adjusted through the associated reference video character, and then the second candidate interest segment is located through the adjusted video character to obtain the second candidate interest segment, thereby improving the reliability of locating the second candidate interest segment.

[0120] In a possible implementation, priority analysis can be performed on the priority analysis results of the video characters, and reminders can be displayed, as follows:

[0121] F1. Determine the first priority through the priority analysis result;

[0122] F2. If the first priority is higher than a preset priority, display the first priority video character and the second priority video character.

[0123] Different priority analysis results have their corresponding priorities, so that a first priority corresponding to the second candidate interest segment can be obtained.

[0124] The preset priority is set by experience value or historical data. If the first priority is higher than the preset priority, it can be explained that the action of the second video character obtained by adjusting the first priority video character may have a large error, and manual verification is required at this time, and subsequent derivation and modification and adjustment actions are performed.

[0125] In this example, after the first priority is higher than the preset priority, the first priority video character and the second priority video character are displayed, so that the user can view and verify them, and make subsequent adjustments, etc., which improves the reliability of processing.

[0126] It should be understood that the size of the sequence numbers of the steps in the above embodiments does not mean the order of response. The response order of each process should be derived from its script and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0127] In another embodiment, a short video playback recommendation device based on big data is provided, and the short video playback recommendation device based on big data corresponds to the short video playback recommendation method based on big data in the above embodiment. Figure 3 As shown, the short video playback recommendation device based on big data includes a collection module 301, a first output module 302, a second output module 303, a third output module 304 and a positioning interception module 305. Each script segment is described in detail as follows:

[0128] The acquisition module 301 is used to acquire metadata of the video to be recommended;

[0129] A first output module 302 is used to obtain POI data corresponding to each of the M candidate interest segments in the video to be recommended through the metadata;

[0130] The second output module 303 is used to obtain user history record data of each candidate interest segment in the M candidate interest segments through the POI data of each candidate interest segment in the M candidate interest segments;

[0131] The third output module 304 is used to obtain a corresponding interest segment through the user history record data of each candidate interest segment in the M segments of the candidate interest segments and the segment location request corresponding to each candidate interest segment in the M segments of the candidate interest segments, so as to obtain a target interest segment;

[0132] The positioning and interception module 305 is used to locate and intercept the target interest segment to obtain the video to be recommended.

[0133] In a possible implementation, the third output module 304 is used to:

[0134] Collecting historical viewing records corresponding to each candidate interest segment in the M candidate interest segments through user history record data of each candidate interest segment in the M candidate interest segments;

[0135] Performing stop viewing time classification processing on the M candidate interest segments to obtain P stop viewing POI lists, wherein each stop viewing POI list in the P stop viewing POI lists has a different stop viewing time, and the candidate interest segments in the stop viewing POI lists have the same stop viewing time;

[0136] According to the corresponding response order of the stop viewing time of the P stop viewing POI lists, the corresponding historical viewing records of the candidate interest segments in the stop viewing POI lists are run in the corresponding element library of the P stop viewing POI lists and the corresponding segment positioning request is responded to to obtain the target interest segment.

[0137] In a possible implementation, before running the corresponding historical viewing records of the candidate interest segments in the stop viewing POI list in the element library corresponding to the P stop viewing POI list and responding to the corresponding segment positioning request to obtain the target interest segment according to the corresponding response order of the stop viewing duration of the P stop viewing POI list, the short video playback recommendation device based on big data is also used to:

[0138] Collecting video access data of historical viewing records corresponding to each candidate interest segment in the M candidate interest segments;

[0139] If the video access frequency indicated by the video access data of the historical viewing record corresponding to the first candidate interest segment among the M candidate interest segments is lower than the video access frequency threshold indicated by the preset video access data, sending history request data to the history server, the history request data including the preset video access data corresponding to the first candidate interest segment;

[0140] Receiving historical viewing records corresponding to preset video access data sent by the history server;

[0141] The historical viewing record corresponding to the preset video access data is obtained as the historical viewing record corresponding to the first candidate interest segment.

[0142] In a possible implementation, the short video playback recommendation device based on big data is also used for:

[0143] Collecting request response results corresponding to each candidate interest segment in the M candidate interest segments;

[0144] Obtaining corresponding response status data of each candidate interest segment in the M segments of candidate interest segments through the corresponding request response result of each candidate interest segment in the M segments of candidate interest segments;

[0145] Obtaining a second candidate interest segment, where the second candidate interest segment is a candidate interest segment whose response status indicated by the response status data is a priority status among the M candidate interest segments;

[0146] collecting a second response parameter of the second candidate interest segment;

[0147] Determining a priority response factor of the second candidate interest segment by using the second response parameter;

[0148] Determining the priority level of the second candidate interest segment according to the priority response factor;

[0149] The priority is displayed.

[0150] In a possible implementation, the short video playback recommendation device based on big data is also used for:

[0151] If the priority factor indicated by the priority response factor is video plot priority, then collecting current video plot status data;

[0152] If the current video plot status data is the priority video plot status data, then displaying the current video plot status data;

[0153] Receiving video plot data corresponding to the second candidate interest segment input by a candidate user;

[0154] The second candidate interest segment is obtained through the historical viewing record corresponding to the second candidate interest segment, the segment location request and the video plot data.

[0155] In a possible implementation, the short video playback recommendation device based on big data is also used for:

[0156] If the priority factor indicated by the priority response factor is video character priority, collecting a first priority video character from the first candidate video characters of the second candidate interest segment by using the second response parameter;

[0157] Performing a priority analysis on the first priority video character to obtain a priority analysis result;

[0158] Adjusting the first priority video character according to the priority analysis result to obtain a second video character;

[0159] Replacing the first priority video character in the first candidate video character corresponding to the second candidate interest segment with the second video character to obtain a second candidate video character;

[0160] The second candidate interest segment is obtained through the second candidate video character, the historical viewing record corresponding to the second candidate interest segment, and the segment positioning request.

[0161] In a possible implementation, the short video playback recommendation device based on big data is also used for:

[0162] Determine a first priority through the priority analysis result;

[0163] If the first priority is higher than the preset priority, the first priority video character and the second priority video character are displayed.

[0164] In this example, by collecting metadata of the video to be recommended, the corresponding POI data of each candidate interest segment in the M candidate interest segments in the video to be recommended is obtained through the metadata, the user history record data of each candidate interest segment in the M candidate interest segments is obtained through the POI data of each candidate interest segment in the M candidate interest segments, the corresponding interest segment is obtained through the user history record data of each candidate interest segment in the M candidate interest segments and the corresponding segment positioning request of each candidate interest segment in the M candidate interest segments, so as to obtain the target interest segment, and the target interest segment is located and intercepted to obtain the video to be recommended. In the present invention, in the positioning of video data for the short video playback platform, the corresponding candidate interest segment can be obtained by collecting the corresponding user history record data of each candidate interest segment in the video to be recommended and the corresponding segment positioning request, and finally the video to be recommended can be obtained by locating and intercepting multiple candidate interest segments, so as to quickly recommend the candidate interest segment without manually collecting the viewing history corresponding to the candidate interest segment, thereby improving the efficiency of locating the candidate interest segment, and further improving the efficiency of locating the video.

[0165] For the specific definition of the short video playback recommendation device based on big data, please refer to the definition of the short video playback recommendation method based on big data above, which will not be repeated here. Each segment in the above-mentioned short video playback recommendation device based on big data can be implemented in whole or in part through software, hardware and its positioning interception. The above-mentioned segments can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and respond to the corresponding operations of each of the above segments.

[0166] In one embodiment, a computer device is provided, which may be a server, and includes a processor, a memory, a video plot interface, and a database connected via a device bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating device, a computer program, and a database. The internal memory provides an environment for the operation of the operating device and the computer program in the non-volatile storage medium. The video plot interface of the computer device is used to communicate with an external client via a video plot connection. When the computer program is responded to by the processor, it implements a script or step on the server side of a short video playback recommendation method based on big data.

[0167] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps in response to the computer program:

[0168] Collect metadata of the video to be recommended;

[0169] Obtaining POI data corresponding to each of the M candidate interest segments in the video to be recommended through the metadata;

[0170] Obtaining user history record data of each candidate interest segment in the M candidate interest segments through the POI data of each candidate interest segment in the M candidate interest segments;

[0171] Obtain a corresponding interest segment through user history record data of each candidate interest segment in the M segments and a segment location request corresponding to each candidate interest segment in the M segments, so as to obtain a target interest segment;

[0172] The target interest segment is located and intercepted to obtain the video to be recommended.

[0173] In this example, by collecting metadata of the video to be recommended, the corresponding POI data of each candidate interest segment in the M candidate interest segments in the video to be recommended is obtained through the metadata, the user history record data of each candidate interest segment in the M candidate interest segments is obtained through the POI data of each candidate interest segment in the M candidate interest segments, the corresponding interest segment is obtained through the user history record data of each candidate interest segment in the M candidate interest segments and the corresponding segment positioning request of each candidate interest segment in the M candidate interest segments, so as to obtain the target interest segment, and the target interest segment is located and intercepted to obtain the video to be recommended. In the present invention, in the positioning of video data for the short video playback platform, the corresponding candidate interest segment can be obtained by collecting the corresponding user history record data of each candidate interest segment in the video to be recommended and the corresponding segment positioning request, and finally the video to be recommended can be obtained by locating and intercepting multiple candidate interest segments, so as to quickly recommend the candidate interest segment without manually collecting the viewing history corresponding to the candidate interest segment, thereby improving the efficiency of locating the candidate interest segment, and further improving the efficiency of locating the video.

[0174] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is responded to by a processor, the following steps are implemented:

[0175] Collect metadata of the video to be recommended;

[0176] Obtaining POI data corresponding to each of the M candidate interest segments in the video to be recommended through the metadata;

[0177] Obtaining user history record data of each candidate interest segment in the M candidate interest segments through the POI data of each candidate interest segment in the M candidate interest segments;

[0178] Obtain a corresponding interest segment through user history record data of each candidate interest segment in the M segments and a segment location request corresponding to each candidate interest segment in the M segments, so as to obtain a target interest segment;

[0179] The target interest segment is located and intercepted to obtain the video to be recommended.

[0180] In this example, by collecting metadata of the video to be recommended, the corresponding POI data of each candidate interest segment in the M candidate interest segments in the video to be recommended is obtained through the metadata, the user history record data of each candidate interest segment in the M candidate interest segments is obtained through the POI data of each candidate interest segment in the M candidate interest segments, the corresponding interest segment is obtained through the user history record data of each candidate interest segment in the M candidate interest segments and the corresponding segment positioning request of each candidate interest segment in the M candidate interest segments, so as to obtain the target interest segment, and the target interest segment is located and intercepted to obtain the video to be recommended. In the present invention, in the positioning of video data for the short video playback platform, the corresponding candidate interest segment can be obtained by collecting the corresponding user history record data of each candidate interest segment in the video to be recommended and the corresponding segment positioning request, and finally the video to be recommended can be obtained by locating and intercepting multiple candidate interest segments, so as to quickly recommend the candidate interest segment without manually collecting the viewing history corresponding to the candidate interest segment, thereby improving the efficiency of locating the candidate interest segment, and further improving the efficiency of locating the video.

[0181] It should be noted that the above-mentioned scripts or steps that can be implemented by computer-readable storage media or computer devices can refer to the relevant descriptions of the server side and the client side in the aforementioned method embodiment. In order to avoid repetition, they will not be described one by one here.

[0182] In the several embodiments provided in the present application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of modules is only a logical division, and there may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another device, or some features can be ignored or not responded. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules, which can be electrical, mechanical or other forms.

[0183] In addition, each script module in each embodiment of the present application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above integrated modules can be implemented in the form of hardware or in the form of software script modules. The above is only an implementation method of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the specification and drawings of this application, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present application.

[0184] The specific implementation methods of the invention are described in detail above, but they are only examples, and the present application is not limited to the specific implementation methods described above. For those skilled in the art, any equivalent modification or substitution of the invention is also within the scope of the present application, and therefore, the equalization, modification, and improvement made without departing from the spirit and principle of the present application should be included in the scope of the present application.

Claims

1. A short video playback recommendation method based on big data, characterized in that: The method comprises: Collect metadata of the video to be recommended; Obtaining POI data corresponding to each of the M candidate interest segments in the video to be recommended through the metadata; Obtaining user history record data of each candidate interest segment in the M candidate interest segments through the POI data of each candidate interest segment in the M candidate interest segments; Obtain a corresponding interest segment through user history record data of each candidate interest segment in the M segments and a segment location request corresponding to each candidate interest segment in the M segments, so as to obtain a target interest segment; Positioning and intercepting the target interest segment to obtain the video to be recommended; The method further comprises: Collecting request response results corresponding to each candidate interest segment in the M candidate interest segments; Obtaining corresponding response status data of each candidate interest segment in the M segments of candidate interest segments through the corresponding request response result of each candidate interest segment in the M segments of candidate interest segments; Obtaining a second candidate interest segment, where the second candidate interest segment is a candidate interest segment whose response status indicated by the response status data is a priority status among the M candidate interest segments; collecting a second response parameter of the second candidate interest segment; Determining a priority response factor of the second candidate interest segment by using the second response parameter; Determining the priority of the second candidate interest segment according to the priority response factor; demonstrating said priorities; After deriving the priority response factor of the second candidate interest segment through the second response parameter, the method further includes: If the priority factor indicated by the priority response factor is video plot priority, then collecting current video plot status data; If the current video plot status data is the priority video plot status data, then displaying the current video plot status data; Receiving video plot data corresponding to the second candidate interest segment input by a candidate user; Determine the second candidate interest segment through the historical viewing record corresponding to the second candidate interest segment, the segment location request and the video plot data; After deriving the priority response factor of the second candidate interest segment through the second response parameter, the method further includes: If the priority factor indicated by the priority response factor is video character priority, collecting a first priority video character from the first candidate video characters of the second candidate interest segment by using the second response parameter; Performing a priority analysis on the first priority video character to obtain a priority analysis result; Adjusting the first priority video character according to the priority analysis result to obtain a second video character; The method further comprises: Determine a first priority through the priority analysis result; If the first priority is higher than the preset priority, displaying the first priority video character and the second video character; otherwise, replacing the first priority video character in the first candidate video character corresponding to the second candidate interest segment with the second video character to obtain the second candidate video character; The second candidate interest segment is obtained through the second candidate video character, the historical viewing record corresponding to the second candidate interest segment, and the segment positioning request.

2. The short video playback recommendation method based on big data according to claim 1 is characterized in that: The step of obtaining a corresponding interest segment through user history record data of each candidate interest segment in the M candidate interest segments and a corresponding segment location request to obtain a target interest segment includes: Collecting historical viewing records corresponding to each candidate interest segment in the M candidate interest segments through user history record data of each candidate interest segment in the M candidate interest segments; Performing stop viewing time classification processing on the M candidate interest segments to obtain P stop viewing POI lists, wherein each stop viewing POI list in the P stop viewing POI lists has a different stop viewing time, and the candidate interest segments in the stop viewing POI lists have the same stop viewing time; According to the corresponding response order of the stop viewing time of the P stop viewing POI lists, the corresponding historical viewing records of the candidate interest segments in the stop viewing POI lists are run in the corresponding element library of the P stop viewing POI lists and the corresponding segment positioning request is responded to to obtain the target interest segment.

3. The short video playback recommendation method based on big data according to claim 2 is characterized in that: Before executing the corresponding historical viewing records of the candidate interest segments in the stop viewing POI list in the element library corresponding to the P stop viewing POI list according to the response order corresponding to the stop viewing duration of the P stop viewing POI list and responding to the corresponding segment positioning request to obtain the target interest segment, the method further includes: Collecting video access data of historical viewing records corresponding to each candidate interest segment in the M candidate interest segments; If the video access frequency indicated by the video access data of the historical viewing record corresponding to the first candidate interest segment among the M candidate interest segments is lower than the video access frequency threshold indicated by the preset video access data, sending history request data to the history server, the history request data including the preset video access data corresponding to the first candidate interest segment; Receiving historical viewing records corresponding to preset video access data sent by the history server; The historical viewing record corresponding to the preset video access data is obtained as the historical viewing record corresponding to the first candidate interest segment.

4. A short video playback recommendation device based on big data, characterized in that: The device comprises: A collection module, used to collect metadata of the video to be recommended; A first output module is used to obtain POI data corresponding to each of the M candidate interest segments in the video to be recommended through the metadata; A second output module is used to obtain user history record data of each candidate interest segment in the M candidate interest segments through POI data of each candidate interest segment in the M candidate interest segments; A third output module is used to obtain a corresponding interest segment through user history record data of each candidate interest segment in the M segments and a segment location request corresponding to each candidate interest segment in the M segments, so as to obtain a target interest segment; A positioning and interception module, used for positioning and intercepting the target interest segment to obtain the video to be recommended; The short video playback recommendation device based on big data is used to execute the short video playback recommendation method based on big data as described in any one of claims 1-3.

Images