Research on automatic labeling of imbalanced texts of customer complaints based on text enhancement and layer-by-layer semantic matching

Due to its potential impact on business efficiency, automated customer complaint labeling and classification are of great importance for management decision making and business applications. The majority of the current research on automated labeling uses large and well-balanced datasets. However, customer complaint labels are hierarchical in structure, with many labels at the lowest hierarchy level. Relying on lower-level labels leads to small and imbalanced samples, thus rendering the current automatic labeling practices inapplicable to customer complaints. This article proposes an automatic labeling model incorporating the BERT and word2vec methods. The model is validated on electric utility customer complaint data. Within the model, the BERT method serves to obtain shallow text tags. Furthermore, text enhancement is used to mitigate the problem of imbalanced samples that emerge when the number of labels is large. Finally, the word2vec model is utilized for deep text analysis. Experiments demonstrate the proposed model's efficiency in automating customer complaint labeling. Consequently, the proposed model supports enterprises in improving their service quality while simultaneously reducing labor costs.

Research on Automatic Labelling of Imbalanced Texts of Customer Complaints Based on Text Enhancement and Layer-By-Layer Semantic Matching

[Purpose/meaning] Due to its potential impact on business efficiency, automated customer complaint labeling and classification are of great importance for management decisions and business-level applications. The majority of the current research on automated labeling uses large and well-balanced datasets. However, customer complaints' labels are hierarchical in structure, with many labels at the lowest hierarchy level. Relying on lower-level labels leads to small and imbalanced samples, thus rendering the current automatic labeling practices not applicable to customer complaints. [Methodology/process] This article proposes an automatic labeling model incorporating the BERT and Word2Vec methods. The model is validated on electric utility customer complaints data. Within the model, the BERT method serves to obtain shallow-level text tags. Further, text enhancement is used to mitigate the problem of uneven samples that emerges when the number of labels is large. Finally, the Word2Vec model is utilized for the deep-level text analysis. [Findings/conclusions] The experiments demonstrate the proposed model's efficiency in automating customer complaint labeling. Consequently, the proposed model supports the enterprises in improving their service quality while simultaneously reducing labor costs.

Effects of Text Enhancement, Identical Prescription-Package Names, Visual Cues, and Verbal Provocation on Visual Searches of Look-Alike Drug Names: A Simulation and Eye-Tracking Study

Objective Simulation and eye tracking were used to examine the effects of text enhancement, identical prescription-package names, visual cues, and verbal provocation on visual searches of look-alike drug names. Background Look-alike drug names can cause confusion and medication errors, which jeopardize patient safety. The effectiveness of many strategies that may prevent these problems requires evaluation. Method We conducted two experiments that were based on a four-way, repeated-measures design. The within-subject factors were text enhancement, identical prescription-package names, visual cues, and verbal provocation. In Experiment 1, 40 nurses searched for and selected a target drug from an array of drug packages on a pharmacy shelf mock-up. In Experiment 2, the eye movements of another 40 nurses were tracked while they performed a computer-based drug search task. Results Text enhancement had no significant effect on the drug search. Nurses selected the target drugs more quickly and easily when the prescriptions and drug packages shared identical drug name formats. The use of a visual cue to direct nurses’ attention facilitated their visual searches and improved their eye gaze behaviors. The nurses reported greater mental effort if they were provoked verbally during the drug search. Conclusion Efficient and practical strategies should be adopted for designs that facilitate accurate drug search. Among these strategies are using identical name appearances on drug prescriptions and packages, using a visual cue to direct nurses’ attention, and avoiding rushing nurses while they are concentrating. Application The findings aim to inspire recommendations for work system designs that will improve the visual search of look-alike drug names.