Skip to main content

Table 1 Features of the destination prediction model.

From: Forecasting worldwide empty container availability with machine learning techniques

Feature

Description

Type

Unique values

Origin

Location of the last event of the container. This can either be a specific port or the country were the event occurred

Categorical

287

Container type

String-based indicator of the type of the container, e.g. FORTY-DRY-CONTAINER

Categorical

23

Carrier

String-based indicator of the carrier of the container, e.g. MSC

Categorical

60

Days passed

Number of days passed between the last event of the container and the date the forecast is made

Numeric

N.A

Event type

String-based indicator of the type of the last event, either dispatch, loading, discharge, or return

Categorical

4