Nowcasting of selected imports and exports of Bangladesh: Comparison among traditional time series model and machine learning models

Md Moyazzem Hossain, Faruq Abdulla, Azizur Rahman

Research output: Book chapter/Published conference paperChapter (peer-reviewed)peer-review

1 Citation (Scopus)

Abstract

Nowcasting, a combination of “now” and “forecast”, is the estimation of a target variable's current state, or a close approximation of it, either forwards or backwards in time, utilizing information that is available in a more timely manner. It has a wide range of applications, all of which attempt to supplement and assist users' decisions. This research applied auto-regressive moving average, neural network models, and support vector regression technique for modelling and nowcasting selected imports and exports in Bangladesh considering annual data from 1976–2020. The findings revealed that support vector models had superior performance compared to the other models considered in this study. In economic growth modelling and nowcasting purpose, the author recommends using the machine learning methodologies. It also suggests that the results be compared to classic econometric and time series models considering other variables with longer data periods.
Original languageEnglish
Title of host publicationComputational statistical methodologies and modelling for Artificial Intelligence
EditorsPriyanka Harjule, Azizur Rahman, Basant Agarwal, Vinita Tiwari
Place of PublicationBoca Raton
PublisherCRC Press
Chapter6
Pages159-177
Number of pages19
Edition1st
ISBN (Electronic)9781003253051
ISBN (Print)9781032170800
DOIs
Publication statusPublished - 2023

Fingerprint

Dive into the research topics of 'Nowcasting of selected imports and exports of Bangladesh: Comparison among traditional time series model and machine learning models'. Together they form a unique fingerprint.

Cite this