Grouping Indonesian Province Farmers’ Term of Trade Using Dynamic Time Warping

<p>This study employs dynamic time warping (DTW) to analyze the farmer’s terms of trade (FTT) across 34 provinces in Indonesia, aiming to identify patterns and cluster similarities in time series data. DTW is recognized for its effectiveness in measuring flexible similarities under time distortions, making it particularly suitable for time series classification across various fields. The FTT is utilized to assess farmers' purchasing power by comparing the prices they receive for their products to the prices they pay for goods and services. K-Medoid clustering techniques were applied to group provinces based on their DTW distances, revealing three distinct clusters. The silhouette score indicates that three clusters as the optimum cluster for the FTT data. The findings show that the first and third clusters have low mean of FTT and the second cluster has the highest mean FTT. These indicates disparities in farmers’ income and purchasing power across regions where the government needs to enhance agricultural strategies and improve economic conditions for farmers in the first and third clusters.</p><p><strong><em>Keywords</em></strong><strong><em>: </em></strong><em>Clustering; Dynamic Time Warping; Farmers Term of Trade; K-Medoid</em><em>.</em></p>