A Conceptual Model for Coeval Time Estimation of Virtual Reference Service

Abstract

The objective of this paper is to propose a conceptual model of a virtual reference service (VRS) time estimation system that can estimate and display the time required to complete a transaction. The system will start estimating the duration before initiation of a transaction and update the estimated duration with the changing complexity of the transaction. Literature on industrial time estimation techniques and artificial-intelligence based VRS programs are reviewed to identify the gap. Four flowcharts are constructed with the aid of machine learning techniques like random forest regression and speech recognition, and natural language processing techniques like intent and entity recognition, and weighting. The pre-estimating, estimating, and post-estimating stages of the system are vividly explained with an example to enlighten the time-estimation process. The limitation of the paper is that the system is not practically developed and tested. However, developing such a system will help in transparent and unbiased transaction time estimation and assist the library professionals to manage queue time and maintain consistent timeliness. The estimated time may be used as a benchmark for evaluating different aspects of VRS. Providing quality service within the estimated time may increase the reliability and loyalty of the patrons toward the library professionals.