一种基于文本相似度模型的人格分析方法
首发时间:2021-01-14
摘要:基于文本信息进行人格分析拥有着广泛的应用领域,并且在近些年获得了极大的学术关注。尽管现有的基于文本进行人格分析的方法已经取得了很大的进步,然而在效果、特征提取、数据需求方面都还存在着一定的缺陷。本文提出一种基于文本相似度模型的人格分析方法,该方法放弃了传统模型直接从用户数据中获取文本特征到人格特征的映射关系思路,转为训练模型学习具有不同人格特征的人物之间的语义相似度差异的问题,进而降低了模型学习难度,减少了模型对数据量的需求,提升了模型的预测效果。与六种著名方法在两个人格数据集上的对比试验证明,本方法取得了最优效果,具有有效性与优越性。
For information in English, please click here
A Personality Analysis Method Based on Text Similarity Model
Abstract:Personality analysis based on text information has a wide range of applications and has gained great academic attention in recent years. Although existing personality analysis methods based on texts have made considerable progress, there are some defects in terms of effect, feature extraction and data annotation. In this paper, we propose a personality analysis method based on text similarity. Our method is different from traditional methods that directly obtain the mapping relationship between text features and personality features from user data. Instead, we turn to training models to learn the difference in semantic similarity between people with different personality characteristics, thereby reducing the difficulty of model learning, alleviating the sparse labeled data problem and improving the prediction accuracy of the model.In comparative experiments with six well-known methods on two personality datasets, this method has achieved the best results and proved the effectiveness and superiority of our method.
Keywords: Deep Learning Semantic Similarity Personality Analysis BigFive Personality Traits
基金:
引用
No.****
动态公开评议
共计0人参与
勘误表
一种基于文本相似度模型的人格分析方法
评论
全部评论0/1000