Topic🧑🏽🦳：Data Exploratory Analysis and Probability Prediction of Slot Booking Cancellation for Container Liner Shipping Services
Speaker：Prof. Qiang Meng,Department of Civil and Environmental Engineering, National University of Singapore
Time: 15th Dec. 2021, 10:00
Online Tencent Meeting ID🤸🏿：
https://meeting.tencent.com/dm/ebk9HUhiISH0
ID:773-963-807
Off Line Venue: Lecture Hall at College of Transport and Communications
Abstract：
Although customers book container slots from shipping companies to transport their cargoes, many of container slot bookings are cancelled, and these cancellations result in low loading rates of ships, which has been a pain point of the container liner shipping industry. However, the cancellation of container slot booking is rarely investigated in the literature. Therefore, the pattern of container slot booking cancellation is unclear to either the container shipping industry or academia. To fill this research gap, this study first studies the cancellation behavior of slot bookings for an intercontinental container shipping service. We develop a viable data exploratory analysis model and use the real slot booking data of Asia to US West Coast service for analysis. This study defines the KPIs to indicate cancellation patterns of shipping voyage. We also discuss the key influential factors on cancellation behavior and investigate the direction of influence. After the data exploratory analysis, we estimate the cancellation probability of container slot bookings by considering these influential factors. We further propose a time-to-cancellation model based on proportional hazard regression. We novelty incorporate the regional feature of container shipping market as a frailty term in model to capture the characteristics of container shipping market. The case study shows that the proposed model is effective in predicting the loaded container volume. It also proves that the model with frailty term can effectively reflect the market regionality and thus improve the model performance. We also shed light on the implications of the coefficients of the covariates and a method to rank the cancellation risk of different markets.