CN-120336619-A - Live broadcast data processing method and system

CN120336619ACN 120336619 ACN120336619 ACN 120336619ACN-120336619-A

Abstract

The invention discloses a live broadcast data processing method and system, and relates to the technical field of data processing. According to the live broadcast data processing method, user behavior data in the current period are obtained in real time, data analysis is conducted respectively to obtain interaction indexes and preference indexes in the current period, text information in the current period is obtained at the same time, language analysis processing is conducted to obtain comment emotion indexes in the current period in the live broadcast process, comprehensive analysis is conducted by combining the interaction indexes and the preference indexes to obtain live broadcast effect feedback indexes in the current period in the live broadcast process, and corresponding live broadcast measures are adopted for the next period based on the live broadcast effect feedback indexes in the current period in the live broadcast process, so that live broadcast content and interaction strategies are optimized in real time, adjustment of the live broadcast content and strategies becomes more flexible and accurate, and viscosity and liveness of audiences are improved.

Inventors

SONG QIUFENG

Assignees

JIANGSU FENGFAN TECH CO LTD

Dates

Publication Date: 20250718
Application Date: 20250318
Priority Date: 20250318

Claims (8)

1. A live broadcast data processing method, comprising the steps of: acquiring user behavior data of a current period in a live broadcast process in real time; Respectively carrying out data analysis on user behavior data of the current period in the live broadcast process to obtain an interaction index and a preference index of the current period in the live broadcast process; simultaneously acquiring text information of a current period in a live broadcast process, performing language analysis processing to obtain comment emotion indexes of the current period in the live broadcast process, and performing comprehensive analysis by combining the interaction indexes and preference indexes to obtain a live broadcast effect feedback index of the current period in the live broadcast process; and taking corresponding live measures for the next time period based on the live effect feedback index of the current time period in the live process.
2. The live data processing method of claim 1, wherein the user behavior data includes a comment number value, a praise number value, a share number value, a bullet screen frequency value, a viewer number value, a viewing duration value, a user churn value, a viewing quality value.
3. The live data processing method according to claim 2, wherein the viewing duration value is a mean value of viewing durations of each user of a current period in a live broadcast process, and the viewing quality value is a mean value of resolutions of viewing live broadcast of each user.
4. The live broadcast data processing method according to claim 2, wherein the specific steps of obtaining the interaction index of the current period in the live broadcast process are as follows: Normalizing the comment number value, the praise number value, the sharing number value and the bullet screen frequency value of the current period in the live broadcast process; and comprehensively analyzing based on the comment number value, the praise number value, the sharing number value and the bullet screen frequency value of the current period in the live broadcast process after normalization processing to obtain the interaction index of the current period in the live broadcast process.
5. The live broadcast data processing method according to claim 2, wherein the specific step of obtaining the preference index of the current period in the live broadcast process is as follows: carrying out standardized processing on the value of the viewer, the value of the viewing time length, the value of the user loss rate and the value of the viewing quality in the current period in the live broadcast process; And comprehensively analyzing based on the value of the watching person, the value of the watching duration, the value of the user loss rate and the value of the watching quality in the current period in the live broadcast process after the standardized processing to obtain the preference index of the current period in the live broadcast process.
6. The live broadcast data processing method according to claim 1, wherein the specific steps of obtaining comment emotion indexes of a current period in a live broadcast process are as follows: reading text information of the current period in the live broadcast process, and preprocessing; Inputting the text information of the current period in the pre-processed live broadcast process to a pre-trained natural language processing model for predictive analysis to obtain the emotion score of each comment of the current period in the live broadcast process; and comprehensively analyzing the emotion score of each comment in the current period in the live broadcast process to obtain the comment emotion index in the current period in the live broadcast process.
7. The live data processing method according to claim 1, wherein the specific steps of taking corresponding live action for the next time period based on the live effect feedback index of the current time period in the live process are as follows: judging and analyzing a live broadcast effect feedback index of the current period in the live broadcast process and a preset live broadcast effect feedback index threshold; If the feedback index of the live broadcast effect in the current period in the live broadcast process is lower than a preset feedback index threshold of the live broadcast effect, taking a first direct broadcast measure for the next period; if the live effect feedback index of the current period in the live process is not lower than the preset live effect feedback index threshold, taking a second live measure for the next period.
8. A live data processing system applying the live data processing method of any of claims 1-7, comprising: The data acquisition module is used for acquiring user behavior data of the current period in the live broadcast process in real time; The data analysis module is used for respectively carrying out data analysis on the user behavior data of the current period in the live broadcast process to obtain an interaction index and a preference index of the current period in the live broadcast process; the language comprehensive analysis module is used for simultaneously acquiring text information of the current period in the live broadcast process, carrying out language analysis processing to obtain comment emotion indexes of the current period in the live broadcast process, and carrying out comprehensive analysis by combining the interaction indexes and preference indexes to obtain a live broadcast effect feedback index of the current period in the live broadcast process; and the live broadcast feedback module is used for taking corresponding live broadcast measures for the next time period based on the live broadcast effect feedback index of the current time period in the live broadcast process.

Description

Live broadcast data processing method and system Technical Field The invention relates to the technical field of data processing, in particular to a live broadcast data processing method and system. Background With the continuous development of internet technology, the live broadcast industry has become one of the important ways of digital entertainment and information dissemination. From the original social live platform to the diversified applications such as video conference, online education, live broadcast of electronic commerce and the like, living broadcast has deeply changed the life style and working mode of people, and particularly in recent years, with the continuous progress of live broadcast technology, interactivity and instantaneity are enhanced, and communication between a live broadcast platform and a user is more compact. In the prior art, the traditional live broadcast data processing method generally depends on simple data acquisition and storage, so that potential value behind the data is difficult to deeply mine, and although a large amount of user behavior and interaction data can be collected by many live broadcast platforms, certain hysteresis still exists in the aspects of data processing and analysis, so that accurate user preference and requirement are difficult to acquire in real time. Limitations of the prior art at least include the following problems, in the prior art, interaction data and preference data are separated, effective combination is lacking, user interests are difficult to comprehensively evaluate, so that content is difficult to timely adjust according to real-time audience behaviors and interactions in a live broadcast process, dynamic changes of audience emotions are difficult to capture in real time in the prior art, emotion feedback in the interaction data is ignored, comprehensive effect evaluation and dynamic adjustment are lacking, accurate effect evaluation and content optimization are difficult to perform, and corresponding live broadcast strategy adjustment is difficult to make in real time. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a live broadcast data processing method and system, which solve the problems that interaction data and preference data are separated and audience emotion changes are difficult to capture in real time in the prior art, so that comprehensive evaluation and dynamic adjustment are easy to lack, and further, content and strategies are difficult to accurately optimize. The live broadcast data processing method comprises the following steps of obtaining user behavior data of a current period in a live broadcast process in real time, respectively carrying out data analysis on the user behavior data of the current period in the live broadcast process to obtain an interactive index and a preference index of the current period in the live broadcast process, simultaneously obtaining text information of the current period in the live broadcast process, carrying out language analysis processing to obtain a comment emotion index of the current period in the live broadcast process, carrying out comprehensive analysis by combining the interactive index and the preference index to obtain a live broadcast effect feedback index of the current period in the live broadcast process, and taking corresponding live broadcast measures for the next period based on the live broadcast effect feedback index of the current period in the live broadcast process. Further, the user behavior data comprises comment number value, praise number value, sharing number value, bullet screen frequency value, viewer number value, viewing duration value, user loss value and viewing quality value. Further, the viewing duration value is a mean value of the viewing duration of each user in the current period in the live broadcast process, and the viewing quality value is a mean value of the resolution of each user for viewing the live broadcast. Further, the method comprises the specific steps of carrying out normalization processing on the comment quantity value, the praise value, the sharing number value and the barrage frequency value of the current period in the live broadcast process, and carrying out comprehensive analysis on the comment quantity value, the praise value, the sharing number value and the barrage frequency value of the current period in the live broadcast process after normalization processing to obtain the interaction index of the current period in the live broadcast process. Further, the specific steps of obtaining the preference index of the current period in the live broadcast process are as follows, standardized processing is carried out on the watching person value, the watching time length value, the user loss value and the watching quality value of the current period in the live broadcast process, and comprehensive analysis is carried out on the basis of the watching person value, the watching time le