Abstract

Recommendation Systems focus on providing the consumers with the closest matches of items that suites the tastes of the consumer. The tastes are identified, based either on the history of purchases made by the consumer (content based) or by accounting for the association born out as a result of their comparison with their closest companions. These companions are the ones who shared similar interests to that of the consumer (collaborative filtering). In this paper, we propose Recommendation System for Crops (CRS),that recommends the appropriate crop(s) for plantation in a particular region, taking into consideration the nature of soil, the climatic conditions, the amount of rainfall, the temperature levels and the water stress impact measures of the region under examination. The system has shown 77% prediction accuracy, proved by validation tests conducted against predictions of classifier models, which are built out of data, gathered by processing image datasets (maps) of Indian territory.

Description