A Comparative Study of Multivariate Clustering Algorithms for Climate Variables and Chickpea Production in Dharwad District of Karnataka, India
Nikil Gowda A. M.
Department of Agricultural Statistics, University of Agricultural Sciences, Dharwad, India.
Ashalatha K. V. *
Department of Agricultural Statistics, University of Agricultural Sciences, Dharwad, India.
Megha J.
Department of Agricultural Statistics, University of Agricultural Sciences, Dharwad, India.
*Author to whom correspondence should be addressed.
Abstract
India's agriculture sector holds immense significance in its economy, and this study delves into how different clustering techniques can enhance our comprehension of the interplay between weather patterns and chickpea production in Dharwad district, specifically Density-Based Spatial Clustering of Applications with Noise (DBSCAN), K-Means, and Hierarchical clustering methods to discern underlying patterns within climate variables and chickpea yield data. Leveraging historical weather data spanning an extensive 42-year period and crop yield records were employed to train and validated. The findings illuminate that K-Means clustering consistently outperforms DBSCAN and Hierarchical clustering when evaluated through a variety of validity indices, including the Silhouette Coefficient, Calinski-Harabasz Index, Davies-Bouldin Index, and the determination of the optimal number of clusters. In essence, India's pivotal agricultural sector dynamics and the intricate relationship between climatic factors and chickpea production in Dharwad district are elucidated effectively by the superiority of K-Means clustering in this study.
Keywords: Density-Based spatial clustering of applications with noise, K-Means Clustering, hierarchical clustering, silhouette coefficient, Calinski-Harabasz index, Davies-Bouldin index