CN-121996803-A - Interactive data prediction method and device for multimedia data, storage medium and electronic equipment

CN121996803ACN 121996803 ACN121996803 ACN 121996803ACN-121996803-A

Abstract

The embodiment of the disclosure provides an interactive data prediction method and device for multimedia data, a storage medium and electronic equipment. The method comprises the steps of obtaining to-be-processed multimedia data, determining a target class cluster to which the to-be-processed multimedia data belongs from clustering class clusters of the historical multimedia data based on the to-be-processed multimedia data, wherein the target class cluster is preset with corresponding historical interaction statistical data, the historical interaction statistical data corresponding to the target class cluster are determined based on the historical interaction data of the historical multimedia data in the target class cluster, determining statistical offset data corresponding to the to-be-processed multimedia data based on a trained prediction model, and determining predicted interaction data of the to-be-processed multimedia data based on the historical interaction statistical data and the statistical offset data corresponding to the target class cluster. And carrying out interactive data prediction on the multimedia data to be processed through a fusion clustering algorithm and a prediction model, and improving the prediction accuracy.

Inventors

ZHAO CHENG

Assignees

北京字跳网络技术有限公司

Dates

Publication Date: 20260508
Application Date: 20241101

Claims (11)

1. An interactive data prediction method for multimedia data, comprising: acquiring multimedia data to be processed; Determining a target class cluster to which the to-be-processed multimedia data belongs in a cluster class cluster of the historical multimedia data based on the to-be-processed multimedia data, wherein the target class cluster is preset with corresponding historical interaction statistical data, and the historical interaction statistical data corresponding to the target class cluster is determined based on the historical interaction data of the historical multimedia data in the target class cluster; Determining statistical offset data corresponding to the multimedia data to be processed based on the trained prediction model; And based on the historical interaction statistical data and the statistical offset data corresponding to the target class cluster, the predicted interaction data of the multimedia data to be processed is truly obtained.
2. The method according to claim 1, wherein the determining, based on the to-be-processed multimedia data, a target class cluster to which the to-be-processed multimedia data belongs in a cluster class cluster of historical multimedia data includes: for any cluster in the cluster clusters, matching the to-be-processed multimedia data with the to-be-processed multimedia data based on cluster sample data of the cluster clusters, and determining target matching data of the to-be-processed multimedia data and the cluster clusters; and determining the target class cluster to which the multimedia data to be processed belongs based on the multimedia data to be processed and the target matching data of each class cluster respectively.
3. The method of claim 2, wherein determining target match data for the multimedia data to be processed and the class cluster based on matching the class cluster sample data for the class cluster with the multimedia data to be processed, comprises: determining content matching data based on the multimedia data to be processed and each of the cluster-like sample data; determining information matching data based on the multi-dimensional information of the multimedia data to be processed and the multi-dimensional information of each cluster-like sample data; And determining target matching data of the to-be-processed multimedia data and the class clusters based on the content matching data and/or the information matching data of the to-be-processed multimedia data and each class cluster sample data respectively.
4. The method of claim 1, wherein the method of determining the cluster type cluster comprises: Acquiring historical multimedia data, and carrying out clustering processing on the historical multimedia data based on the number of the plurality of the class clusters to obtain candidate clustering class clusters corresponding to the number of the plurality of the class clusters respectively; and determining the cluster clusters of the target based on the contour coefficients respectively corresponding to the candidate cluster clusters.
5. The method according to claim 4, wherein the method further comprises: For each cluster in the cluster type clusters, based on the historical interaction data of a plurality of historical multimedia data in the cluster type clusters in a plurality of time periods of a preset playing period, determining the historical interaction statistical data of the cluster type clusters, wherein the historical interaction statistical data of the cluster type clusters represent the interaction data trend of the cluster type clusters in the preset playing period.
6. The method of claim 5, wherein the cluster-like historical interaction statistics include historical interaction statistics corresponding to a plurality of time periods of the preset playing period, respectively; The statistical offset data corresponding to the multimedia data to be processed is the statistical offset data corresponding to the multimedia data to be processed in a target time period; based on the historical interaction statistical data and the statistical offset data corresponding to the target class cluster, the predicted interaction data of the multimedia data to be processed is truly included: and determining predicted interaction data of the to-be-processed multimedia data in the target time period based on the historical interaction statistical data of the target class cluster in the target time period and the statistical offset data of the to-be-processed multimedia data in the target time period.
7. The method of claim 6, wherein the historical interaction statistics comprise at least one statistical indicator, and wherein the statistical offset data comprises offset data corresponding to the at least one statistical indicator, respectively; Based on the historical interaction statistical data of the target class cluster corresponding to the target time period and the statistical offset data of the to-be-processed multimedia data corresponding to the target time period, determining the predicted interaction data of the to-be-processed multimedia data in the target time period comprises the following steps: Determining index data of each statistical index corresponding to the target time period of the multimedia data to be processed based on each statistical index corresponding to the target time period of the target class cluster and offset data corresponding to each statistical index; And actually obtaining the predicted interactive data of the multimedia data to be processed in the target time period based on the index data of the at least one statistical index corresponding to the multimedia data to be processed in the target time period.
8. The method of claim 6, wherein determining statistical offset data corresponding to the multimedia data to be processed based on the trained predictive model comprises: Acquiring multi-dimensional information of the multimedia data to be processed, time information of the target time period and interactive data of at least one time period before the target time period in a preset playing period; And processing the multidimensional information of the multimedia data to be processed, the time information of the target time period and the interactive data of at least one time period before the target time period in the preset playing period based on the trained prediction model, and determining the statistical offset data corresponding to the multimedia data to be processed in the target time period.
9. An interactive data prediction apparatus for multimedia data, comprising: The data acquisition module is used for acquiring the multimedia data to be processed; The clustering module is used for determining a target class cluster to which the to-be-processed multimedia data belongs from clustering class clusters of the historical multimedia data based on the to-be-processed multimedia data, wherein the target class cluster is preset with corresponding historical interaction statistical data, and the historical interaction statistical data corresponding to the target class cluster is determined based on the historical interaction data of the historical multimedia data in the target class cluster; the offset data determining module is used for determining statistical offset data corresponding to the multimedia data to be processed based on the trained prediction model; and the interaction data prediction module is used for predicting interaction data of the multimedia data to be processed based on the historical interaction statistical data and the statistical offset data corresponding to the target class cluster.
10. An electronic device, the electronic device comprising: One or more processors; Storage means for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the interactive data prediction method of multimedia data as claimed in any one of claims 1 to 8.
11. A storage medium containing computer executable instructions which, when executed by a computer processor, are for performing the interactive data prediction method of multimedia data as claimed in any one of claims 1 to 8.

Description

Interactive data prediction method and device for multimedia data, storage medium and electronic equipment Technical Field The embodiment of the disclosure relates to the technical field of data processing, in particular to an interactive data prediction method and device for multimedia data, a storage medium and electronic equipment. Background The playing life cycle of the multimedia data comprises the processes of birth, gradual watching, continuous watching and declining, and the like, and the interactive data of the multimedia data at different stages of the playing life cycle are different. The interactive data of the multimedia data at the future time is beneficial to accurately pushing the multimedia data to the user. However, the current prediction mode of the interactive data of the multimedia data is single, and the problem of inaccurate prediction of the interactive data exists. Disclosure of Invention The disclosure provides an interactive data prediction method and device for multimedia data, a storage medium and electronic equipment, so as to improve the accuracy of interactive data prediction of the multimedia data. In a first aspect, an embodiment of the present disclosure provides an interactive data prediction method for multimedia data, including: acquiring multimedia data to be processed; Determining a target class cluster to which the to-be-processed multimedia data belongs in a cluster class cluster of the historical multimedia data based on the to-be-processed multimedia data, wherein the target class cluster is preset with corresponding historical interaction statistical data, and the historical interaction statistical data corresponding to the target class cluster is determined based on the historical interaction data of the historical multimedia data in the target class cluster; Determining statistical offset data corresponding to the multimedia data to be processed based on the trained prediction model; And based on the historical interaction statistical data and the statistical offset data corresponding to the target class cluster, the predicted interaction data of the multimedia data to be processed is truly obtained. In a second aspect, an embodiment of the present disclosure further provides an interactive data prediction apparatus for multimedia data, including: The data acquisition module is used for acquiring the multimedia data to be processed; The clustering module is used for determining a target class cluster to which the to-be-processed multimedia data belongs from clustering class clusters of the historical multimedia data based on the to-be-processed multimedia data, wherein the target class cluster is preset with corresponding historical interaction statistical data, and the historical interaction statistical data corresponding to the target class cluster is determined based on the historical interaction data of the historical multimedia data in the target class cluster; the offset data determining module is used for determining statistical offset data corresponding to the multimedia data to be processed based on the trained prediction model; and the interaction data prediction module is used for predicting interaction data of the multimedia data to be processed based on the historical interaction statistical data and the statistical offset data corresponding to the target class cluster. In a third aspect, an embodiment of the present disclosure further provides an electronic device, including: One or more processors; Storage means for storing one or more programs, The one or more programs, when executed by the one or more processors, cause the one or more processors to implement an interactive data prediction method for multimedia data as provided by embodiments of the present disclosure. In a fourth aspect, the disclosed embodiments also provide a storage medium containing computer-executable instructions, which when executed by a computer processor, are for performing an interactive data prediction method for multimedia data as provided by the disclosed embodiments. According to the embodiment of the disclosure, clustering processing is performed on the historical multimedia data to obtain clustering clusters of the historical multimedia data, and clustering processing is performed on the multimedia data to be processed to obtain target clusters to which the multimedia data to be processed belongs, so that historical interaction statistical data corresponding to the target clusters is obtained. And carrying out prediction processing on the multimedia data to be processed through a prediction model to obtain statistical offset data corresponding to the multimedia data to be processed, and correcting historical interaction statistical data corresponding to the target class cluster based on the statistical offset data corresponding to the multimedia data to be processed to obtain predicted interaction data of the multimedia data to be processed. And carrying out in