Forecasting Developments in the Agricultural Sector in Kosovo through Machine Learning

Session

Computer Science and Communication Engineering

Description

The agricultural sector represents a vital pillar for economic development and food security in Kosovo, constantly facing complex challenges that require innovative approaches to strategic planning. This paper provides an in-depth analysis and forecasts for the performance of this sector until 2030, applying advanced data science techniques. Using a comprehensive dataset covering the period 2007–2023, the study integrates key agricultural indicators, macroeconomic variables and climate data to model the dynamics of the sector. The initial phase of the work included a rigorous process of data preparation, moving from their cleaning and transformation, to normalization and selection of the most influential features through principal component analysis, in order to increase the accuracy of the models. At the core of this study lies the application of a hybrid methodology. Sophisticated machine learning models, namely XGBoost and Random Forest, known for their high performance in predictive tasks with complex data, were built and evaluated. In addition, the ARIMA statistical model, specialized in time series analysis, was used to capture trends and seasonality in historical data. By combining these approaches, the study generates robust forecasts for the period 2024–2030. The results of this research provide a solid basis for informed decision-making, supporting public institutions, agricultural organizations and private investors in designing effective policies and adapting to future challenges, such as climate change and the need for a sustainable development of Kosovo agriculture.

Keywords:

Agriculture, Forecasting, Machine Learning, Data Science

Proceedings Editor

Edmond Hajrizi

ISBN

978-9951-982-41-2

Location

UBT Lipjan, Kosovo

Start Date

25-10-2025 9:00 AM

End Date

26-10-2025 6:00 PM

DOI

10.33107/ubt-ic.2025.89

This document is currently not available here.

Share

COinS
 
Oct 25th, 9:00 AM Oct 26th, 6:00 PM

Forecasting Developments in the Agricultural Sector in Kosovo through Machine Learning

UBT Lipjan, Kosovo

The agricultural sector represents a vital pillar for economic development and food security in Kosovo, constantly facing complex challenges that require innovative approaches to strategic planning. This paper provides an in-depth analysis and forecasts for the performance of this sector until 2030, applying advanced data science techniques. Using a comprehensive dataset covering the period 2007–2023, the study integrates key agricultural indicators, macroeconomic variables and climate data to model the dynamics of the sector. The initial phase of the work included a rigorous process of data preparation, moving from their cleaning and transformation, to normalization and selection of the most influential features through principal component analysis, in order to increase the accuracy of the models. At the core of this study lies the application of a hybrid methodology. Sophisticated machine learning models, namely XGBoost and Random Forest, known for their high performance in predictive tasks with complex data, were built and evaluated. In addition, the ARIMA statistical model, specialized in time series analysis, was used to capture trends and seasonality in historical data. By combining these approaches, the study generates robust forecasts for the period 2024–2030. The results of this research provide a solid basis for informed decision-making, supporting public institutions, agricultural organizations and private investors in designing effective policies and adapting to future challenges, such as climate change and the need for a sustainable development of Kosovo agriculture.