An Explainable Artificial Intelligence Approach for Detecting Empathy in Textual Communication

被引:9
作者
Montiel-Vazquez, Edwin Carlos [1 ]
Uresti, Jorge Adolfo Ramirez [1 ]
Loyola-Gonzalez, Octavio [2 ]
机构
[1] Tecnol Monterrey, Sch Engn & Sci, Atizapan 52926, Estado De Mexic, Mexico
[2] Stratesys, Calle Torrelaguna 77, Madrid 28043, Spain
来源
APPLIED SCIENCES-BASEL | 2022年 / 12卷 / 19期
关键词
empathy; natural language processing; pattern-based classification; affective computing; databases; PATTERN-BASED CLASSIFICATION; STATISTICAL COMPARISONS; CLASSIFIERS; EVOLUTIONARY; QUOTIENT; TESTS;
D O I
10.3390/app12199407
中图分类号
O6 [化学];
学科分类号
0703 ;
摘要
Empathy is a necessary component of human communication. However, it has been largely ignored in favor of other concepts such as emotion and feeling in Affective computing. Research that has been carried out regarding empathy in computer science lacks a method of measuring empathy based on psychological research. Likewise, it does not present an avenue for expanding knowledge regarding this concept. We provide a comprehensive study on the nature of empathy and a method for detecting it in textual communication. We measured empathy present in conversations from a database through volunteers and psychological research. Subsequently, we made use of a pattern-based classification algorithm to predict the Empathy levels in each conversation. Our research contributions are: the Empathy score, a metric for measuring empathy in texts; Empathetic Conversations, a database containing conversations with their respective Empathy score; and our results. We show that an explicative pattern-based approach (PBC4cip) is, to date, the best approach for detecting empathy in texts. This is by measuring performance in both nominal and ordinal metrics. We found a statistically significant difference in performance for our approach and other algorithms with lower performance. In addition, we show the advantages of interpretability by our model in contrast to other approaches. This is one of the first approaches to measuring empathy in texts, and we expect it to be useful for future research.
引用
收藏
页数:27
相关论文
共 105 条
[41]  
Kumano S., 2011, Proceedings 2011 IEEE International Conference on Automatic Face & Gesture Recognition (FG 2011), P43, DOI 10.1109/FG.2011.5771440
[42]  
Kuncheva L.I., 2005, Combining Pattern Classifiers, Methods and Algorithms
[43]   Measuring empathy: reliability and validity of the Empathy Quotient [J].
Lawrence, EJ ;
Shaw, P ;
Baker, D ;
Baron-Cohen, S ;
David, AS .
PSYCHOLOGICAL MEDICINE, 2004, 34 (05) :911-919
[44]  
Leite I., 2010, Proceedings of the 3rd international workshop on Affective interaction in natural environments - AFFINE'10, P43, DOI DOI 10.1145/1877826.1877839
[45]   Interactive double states emotion cell model for textual dialogue emotion prediction [J].
Li, Dayu ;
Li, Yang ;
Wang, Suge .
KNOWLEDGE-BASED SYSTEMS, 2020, 189
[46]  
Li Q., 2019, arXiv
[47]  
Li YZ, 2017, ADV NEUR IN, V30
[48]   Aspect-based sentiment analysis via affective knowledge enhanced graph convolutional networks [J].
Liang, Bin ;
Su, Hang ;
Gui, Lin ;
Cambria, Erik ;
Xu, Ruifeng .
KNOWLEDGE-BASED SYSTEMS, 2022, 235
[49]  
Likert R., 1932, Archives of psychology, V22, P55
[50]  
Lin Z., 2019, arXiv