An Overview of Weather Forecasting Model for Bangladesh Using Regression Based Machine Learning Techniques

Tracking #: 615-1595


Atik MahabubORCID logo
Al-Zadid Sultan Bin HabibORCID logo

Responsible editor: 

Tobias Kuhn

Submission Type: 

Research Paper


Weather forecasting bears significant impacts in our day-to-day life in every aspect from agricultural perspectives to event management. Weather forecasting becomes an uphill task for countries like Bangladesh where plain lands similarly coin- cide with coastal areas or hill tract areas and weather changes frequently. Weather forecasting contains some predictions of key parameters like wind speed, humidity, temperature, and rainfall. Several previous weather forecasting models used the compli- cated mathematical instruments which were rarely accurate. In this paper, regression-based Machine Learning (ML) models have been presented to predict the weather parameters accurately for Bangladesh. A practical application of ML techniques towards environmental numerical modeling has been developed. The raw dataset has been collected from Bangladesh Meteorological Division (BMD) which includes data of wind speed, humidity, temperature and rainfall for the past five years of Bangladesh of several weather stations across the country. Several regression algorithms have been used e.g. Support Vector Regression (SVR), Linear Regression, Bayesian Ridge, Gradient Boosting (GB), Extreme Gradient Boosting (XGBoost), Category Boosting (CatBoost), Adaptive Boosting (AdaBoost), k-Nearest Neighbors (KNN) and Decision Tree Regressor (DTR). The output of regression techniques has been compared with the existing forecast-based models which shows that ML-based models are more accurate than conventional methods.



  • Reviewed

Data repository URLs: 

Date of Submission: 

Friday, November 1, 2019

Date of Decision: 

Monday, December 9, 2019

Nanopublication URLs:



Solicited Reviews:

1 Comment

Meta-Review by Editor

The reviewers pointed out a number of major shortcomings, including with respect to presentation, background, and methodology. Therefore, this paper cannot be accepted in its current form.

Tobias Kuhn (