中国科学院心理研究所机构知识库

近年来,随着计算机自然语言处理以及机器学习技术的日愈成熟,利用网络行为预测用户的心理特征逐渐成为跨学科的研究热点,一些学者也随之开始研究利用人工智能方法建立文学人物心理预测模型。目前的文学智能分析使用微博数据建立的预测模型来对文学人物进行分析,这与文学作品中的场景存在差异。本文将迁移学习引入文学智能分析,针对英国文学家毛姆笔下的文学人物的心理特征进行预测,结果发现迁移学习模型使文学人物的心理预测效果有所提升,表明了迁移学习在文学人物心理分析模型中的有效性。

In recent years，with the development of computer natural language processing and machine learning technology，the use of network behavior to predict users＇ psychological characteristics has gradually become a focus of interdisciplinary research． The analysis of literary characters by the intelligent analysis system based on ecological recognition has been proved to be effective，but there are still some deficiencies． In this paper，the idea of transfer learning is introduced into the intelligent analysis system of literature． Taking the literary characters described by the English writer William Somerset Maugham as an example，the prediction effect is found to be improved to a certain extent，thus proving the effectiveness of transfer learning in the psychological analysis model of literary characters．

国家社会科学基金重点项目(17AZD041)