This work aims to enhance the classification of robot failures using natural language processing models and statistical analysis. Utilizing a coded dataset of 8,306 online customer reviews of robotic vacuum cleaners, we applied RoBERTa, to assess each review's positivity, negativity, and neutrality. Analyses showed that failure category (technical, interaction, and service), review length, and star rating impacted the sentiments. The impact of technical problems was the greatest. We discuss how sentiment analysis contributes to the understanding of perceived robot failure severity.
HijaziR.JerniganR. (2009). Modeling compositional data using Dirichlet regression models. Journal of Applied Probability & Statistics, 4(1), 77–91.
2.
HonigS.BartalA.ParmetY.Oron-GiladT. (2022). Using online customer reviews to classify, predict, and learn about domestic robot failures. International Journal of Social Robotics, 14(6), 1389–1404. https://doi.org/10.1007/s12369-022-00929-3
3.
HonigS.Oron-GiladT. (2022). How user presence impacts perceptions and operation routines of robotic vacuum cleaners: A ‘stay at home’ experiment. In BlackN. L.NeumannW. P.NoyI. (Eds.), Proceedings of the 21st Congress of the International Ergonomics Association (IEA 2021) (Lecture Notes in Networks and Systems, Vol. 223, pp. 275–282). Springer. https://doi.org/10.1007/978-3-030-74614-8_34
SannR.LaiP. (2020). Understanding homophily of service failure within the hotel guest cycle: Applying NLP-aspect-based sentiment analysis to the hospitality industry. International Journal of Hospitality Management, 91, 102678. https://doi.org/10.1016/j.ijhm.2020.102678
6.
XuC.ZhengX.YangF. (2023). Examining the effects of negative emotions on review helpfulness: The moderating role of product price. Computers in Human Behavior, 139, 107501. https://doi.org/10.1016/j.chb.2022.107501
