Media content processing method and device, equipment, readable storage medium and product

The content recommendation model obtained by data augmentation and iterative training based on user historical interaction operations is solved, and more efficient and accurate content recommendation is achieved.

CN119669567APending Publication Date: 2025-03-21DOUYIN VISION CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202411748405.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-29
Publication Date
2025-03-21

AI Technical Summary

Technical Problem

The existing content recommendation method has a large amount of calculation, resulting in low content recommendation efficiency, and a large proportion of high-hot content in recommended content streams, which may be biased from the user's browsing habits, resulting in inaccurate recommended content.

Method used

Data augmentation is carried out by obtaining the media content of interactive operations triggered by user history on the terminal device, building a training data set, and iteratively training the training data set to obtain a content recommendation model. Then, based on the content recommendation model, the predicted content sequence that the user may be interested in is determined in the recall dataset.

Benefits of technology

It improves the efficiency of content recommendation, ensures that the recommended content is more in line with the user's browsing habits, reduces the deviation of highly popular content, and improves the diversity and accuracy of recommended content.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The embodiment of the invention provides a media content processing method and device, equipment, a readable storage medium and a product, and the method comprises the steps: obtaining a preset training data set, the training data set comprises a plurality of training data pairs associated with terminal equipment, and each training data pair comprises the identification information of the terminal equipment and a media content sequence, the media content sequence comprises a first preset number of training media contents obtained by expanding a second preset number of target media contents triggering a preset interaction operation on the terminal equipment based on user history, and the training media contents are arranged according to a random sequence; and performing iterative training on the to-be-trained model through the training data set to obtain a content recommendation model. Therefore, the predicted content sequence can be determined in the recall data set based on the content recommendation model, and content recommendation operation is performed based on the predicted content sequence. Based on the content recommendation model, the prediction content sequence which is more fit with the browsing habit of the user and has more diversified content is determined, and the content recommendation accuracy is improved.
Need to check novelty before this filing date? Find Prior Art