Multivariate Prediction of Package Delivery Time

点击翻译按钮获取中文摘要

Delivery of packages is a critical concept for e-commerce and many online stores that deliver merchandise and**goods to consumers. The expected time of delivery is an important factor for vendors and customers alike. The**time of package delivery often depends on a variety of parameters, ranging from the number of packages to be**delivered, the availability of staff and resources, and the distance of the destination to the warehouse, to weather**and road conditions. As a result, exact prediction of time of delivery is often difficult and subject to change based**on these highly variable parameters. This project aims to address this problem and develop machine learning**models based on existing multi-modal data from package shipments, traffic, and weather data, to predict package**delivery times with high confidence. Innovapost is part of The Canada Post Group of Companies, a multi billion**dollar Enterprise, which develops and oversees the entire technical and analytical IT infrastructure for Canada**Post. This project will allow the company to utilize the developed methods to inform recipients of a window for**day/time of delivery with high confidence and be able to update the estimation with changes in the parameters in**real-time. Such an accurate service will result in higher satisfaction rates by customers, which will result in**financial growth by the company and eventually Canada Post.