Innovación tecnológica en la Agricultura Artesanal: Comparación de enfoques de recomendación para optimizar la comercialización: Technological Innovation in Artisanal Agriculture: Comparison of Recommendation Approaches to Optimize Marketing

Abstract

Introduction: Technological innovation has begun to significantly transform artisanal agriculture, especially in rural areas with limited access to information and communication technologies. In this context, recommender systems emerge as key tools to improve agricultural marketing processes by personalizing suggestions and strategies.
Objective: To compare and evaluate three recommended system approaches based on machine learning, collaborative filtering, and user segmentation to identify the most appropriate method to support agricultural marketing in rural areas with limited resources.
Method: A critical literature review was conducted to analyze the performance of the three recommender system approaches in terms of accuracy, applicability, and process optimization. Results: Hybrid recommender systems, which integrate multiple approaches, demonstrated greater than 80% accuracy and greater adaptability to rural contexts. In contrast, while accurate, systems that rely on cloud computing present difficulties for implementation due to technological infrastructure requirements not available in many rural areas.
Conclusions: Recommender systems based on machine learning, collaborative filtering, and user segmentation offer viable solutions for optimizing agricultural marketing in rural settings. Their implementation contributes to the improvement of productivity and decision making, being an effective alternative to traditional methods. Hybrid models stand out for their flexibility and lower dependence on advanced infrastructure, which makes them a suitable option to support small-scale farmers.

Organizations such as the FAO have emphasized that the use of technology in agriculture can be key to bridging market access gaps and strengthening rural economies. This article compares three different approaches to recommendation systems applied to artisanal agriculture: one based on neural networks and cloud computing, and two hybrid approaches that integrate collaborative filtering, customer segmentation, and implicit data analysis. The analysis shows that hybrid models, due to their flexibility and accuracy, are better suited for contexts with limited technological infrastructure, effectively contributing to process optimization in agricultural environments.