Research on Intelligent Classification of Service Hotline Work Orders Based on Semantic Recognition
Journal Title: Urban Mass Transit - Year 2025, Vol 28, Issue 5
Abstract
[Objective] Traditional rail transit network service hotline mainly relies on manual customer service answering calls, manually filling out work orders and handling classifications. Passenger service staff undertake high-intensity and overloaded service work, while the service quality is difficult to be guaranteed. Therefore, it is necessary to introduce semantic recognition technology based on deep learning to achieve digitalized and intelligent operation management. [Method] The systematic requirements for the business classification of current service hotline and the intelligent classification of the work order are analyzed. The word- segmentation logic of semantic analysis is used and the recognition accuracy is improved by establishing a keyword library. Using such library as a domain dictionary, an intelligent text classification model based on distributed text vector representation and integrating the Transformer self-attention mechanism is constructed. In the proposed intelligent text classification model, the attention focus is more concentrated on the words strongly relevant to the classification task, thus reducing the interference of irrelevant words in the context on the classification results, and texts with different semantics in different contexts can also be dynamically displayed to achieve the classification of passengers′ intentions. On this basis, an intelligent classification system for hotline work order records is built. [Result & Conclusion] The experimental results on the real datasets show that the proposed intelligent text classification model has certain effectiveness and correctness. By adopting a highly modular software system design, the automatic classification of work orders is realized, which can effectively improve the work order response speed and reduce the costs of human and material resources. The proposed intelligent text classification model can improve the overall operation service quality and passenger satisfaction.
Authors and Affiliations
Xiaolei MAO
Keywords
Related Articles
- EP ID EP766711
- DOI 10.16037/j.1007-869x.2025.05.033
- Views 17
- Downloads 0
Analysis of Train Signaling System Safety Distance
[Objective] The current calculation of safety distance in the moving block signaling system does not consider adverse working conditions such as faults and low adhesion coefficients, resulting in poor authenticity and lo...
Performance Analysis of Slurry After Sewage Infiltration in Pipe Jacking Construction of Subway Station Entrances and Exits Project
Objective During pipe jacking construction, aging sewage pipes within the construction area are prone to rupture, leading to sewage infiltration into the thixotropic slurry along the soil body, thus impacting the perform...
Detection System Design and Implementation for Foreign Objects in Automatic Platform Door Gap
Objective The detection of foreign objects in platform door gap is critical to metro operational safety. Therefore, it is essential to develop a new anti-clamping detection system for metro platform door, enhancing the s...
Pantograph-Catenary Relation of Urban Rail Transit Lines under Catenary Iced Condition
［Objective］ Under the condition of ice forming on catenary surface (hereinafter referred to as ′iced condition′), the normal current collection of train pantograph is disrupted. In severe cases, arcing may occur betwee...
Practice and Reflection on Renovation of Shanghai Urban Rail Transit Existing Lines
[Objective] Shanghai is one of the first batch of cities in China launching urban rail transit systems. In recent years, Shanghai has initiated renovation projects for existing rail lines, undertaking many pioneering exp...