Social robot detection data set extension method and system based on active learning

Abstract

The invention discloses a social robot detection data set extension method and system based on active learning. The expansion method comprises the steps: collecting microblog user data, extracting a part of data samples for manual annotation, constructing an initial annotation data set, training a classifier by adopting an active learning method, annotating remaining unannotated data sets, and expanding the data sets according to an annotation result. According to the method, the problem that a reliable large sample data set is lacked in social robot detection research in microblogs is solved,full-automatic data annotation and classification are realized, unnecessary human resource consumption is reduced to a great extent, and the data annotation efficiency is improved to a great extent.

Description

technical field [0001] The invention relates to the technical field of microblog data collection collection, in particular to a method and system for expanding a social robot detection data set based on active learning. Background technique [0002] In recent years, the rapid development of Internet technology has accelerated the widespread popularization of Online Social Networks (OSNs). Online social network is a kind of Internet media widely used by people. As a medium in the process of communication and dissemination of information in contemporary society, its influence has penetrated into all aspects of people's daily life. By registering an account and providing corresponding personal information, users can easily and quickly obtain real-time information and interact with family and friends anytime, anywhere. Online social networks have become an important carrier of information transmission in contemporary society. [0003] With the continuous development of online ...

AI Technical Summary

Problems solved by technology

However, due to the lack of reliable labeled data sets in the existing research on social robot detection in Sina Weibo, it has become a major bottleneck restricting the development of social robot detection technology.

[0006] However, the construction of large-scale datasets labeled with normal users and social robots is currently facing problems such as difficulty in obtaining and high labeling costs.

Images