A stability-oriented protocol for farm typology construction
Open Access
- 20 September 2025
- journal article
- Published by Institute of Eastern European Research and Consulting in Agricultural and Resource Economics: International Scientific E-Journal
- Vol. 11 (3) , 281–312
Abstract
Purpose. This study aims to develop a reliable and stable farm typology using advanced data analysis, clustering stability assessment, and expert input. It also intends to establish a replicable protocol for updating and refining the typology in response to the sector’s dynamic changes. The ultimate result is to support evidence-based planning and more effective agricultural policy targeting. Methodology / approach. The analysis is based on data from 2,756 farms selected to reflect the structural diversity of farming systems across Albania. Five clustering algorithms and five internal validation indices were applied to four dataset versions. A composite Stability Score was introduced to identify the most robust configuration. Expert validation was conducted via focus groups with representatives from government, academia, and rural advisory services. Results. K-means with three clusters achieved the highest composite Stability Score (0.72 on a 0–1 scale) across all tested methods and cluster counts, yielding a typology that is statistically consistent, reproducible, and updateable. The empirical application identified three farm types – Farms with arable crops, Market-oriented farms, and Semi-subsistence family farms – observed in the three study regions (Fier, Korçë, and Kukës). Expert validation confirmed the coherence and applicability of the classification. Finally, after validation, the methodology steps were organised into a protocol, designed to be directly applicable to new data and geographical areas. Originality / scientific novelty. This study introduces a novel protocol with a reproducibility logic and documented pseudocode, allowing for future updates without methodological reinvention. It addresses long-standing concerns in typology construction, such as subjectivity and instability, by implementing a data-driven, multi-method, and expert-validated process that can be adapted across different structural contexts. Practical value / implications. The study presents a replicable and scalable methodology for farm classification, enabling systematic updates over time and improving the alignment of agricultural policies with the structural realities of Albania and other countries facing similar structural challenges. In practical terms, the typology can guide the allocation of IPARD grants, national subsidies, and rural infrastructure programs by aligning financial and technical support with the specific needs of each farm type.Keywords
This publication has 36 references indexed in Scilit:
- Spatially-Clustered Spatial Autoregressive Models with Application to Agricultural Market Concentration in EuropeJournal of Agricultural, Biological and Environmental Statistics, 2025
- A farm typology development cycle: From empirical development through validation, to large-scale organisational deploymentAgricultural Systems, 2024
- Farm typologies for understanding farm systems and improving agricultural policyAgricultural Systems, 2023
- Big data, small explanatory and predictive power: Lessons from random forest modeling of on-farm yield variability and implications for data-driven agronomyField Crops Research, 2023
- Typologies of European farmers: approaches, methods and research gapsRegional Environmental Change, 2022
- Stability estimation for unsupervised clustering: A reviewWIREs Computational Statistics, 2022
- Capturing farm diversity at regional level to up-scale farm level impact assessment of sustainable development optionsAgriculture, Ecosystems & Environment, 2011
- Cluster-wise assessment of cluster stabilityComputational Statistics & Data Analysis, 2007
- Landholder Profiling and Typologies for Natural Resource–Management Policy and Program Support: Potential and ConstraintsEnvironmental Management, 2007
- Silhouettes: A graphical aid to the interpretation and validation of cluster analysisJournal of Computational and Applied Mathematics, 1987