American
International
University-
Bangladesh

Students & Supervisors

Abstract

Bangladesh's economy is unstable because the data isn't always right, the structures aren't strong, and the country can be affected by things that happen in other countries.This makes it very hard to guess how much prices will go up. This study creates a complete machine learning framework that combines different supervised learning methods to make inflation forecasts more accurate, reliable, and understandable in situations where data is unclear or noisy. A systematic data preparation pipeline was employed, including data cleaning, handling of missing values, detection of noisy and inconsistent observations, encoding of the target variable, and normalization of features to preserve data integrity. This study demonstrated that ensemble learning methods, especially the Random Forest, work better than standard regression and single classifiers because they can handle nonlinearities, multicollinearity, and changes in the structure of inflation dynamics. The Random Forest model was the best because it could deal with complicated, nonlinear interactions and noisy data better than the other models. This paper offers a thorough, clear, and data-driven approach for predicting inflation in Bangladesh's economy, combining empirical rigor with ease of understanding. It gives policymakers the tools they need to keep the money stable and make decisions based on facts. This study builds a complete machine learning framework that puts together a lot of supervised learning methods to make inflation predictions more accurate, reliable, and easy to understand when the data is unclear and noisy. The local economy often has fundamental imbalances and data differences that are hard for Ordinary Least Squares (OLS) and other common econometric methods to deal with.The study found that ensemble learning methods, especially the Random Forest, do a much better job than standalone classifiers and standard regression. Explainable research also helps people understand the model better by showing how different factors change the inflation results. With a number of 0.807, the Producer Price Index (PPI) is the most important part. The suggested plan is based on facts and has a clear way to get things done, so lawmakers can trust it. These results can help people make good money decisions and keep the economy safe as it grows

Keywords

Publication Details

• DOI: Not yet Activated
• Type of Publication:
• Conference Name: 2026 IEEE International Conference on Electrical, Computer & Telecommunication Engineering (ICECTE 2026)
• Date of Conference: 29/01/2026 - 29/01/2026
• Venue: Rajshahi-6204, Bangladesh
• Organizer: RUET