Feature fusion based deep neural collaborative filtering model for fertilizer prediction

With the advent of the modern era, deep neural networks have dominated recommender systems, as they can effectively capture complex interactions. Nevertheless, there is still a research gap in fertilizer recommendations based on soil nutrients content. In this work, we propose nutrient-centered deep collaborative filtering technique to determine the required amount of fertilizers for sustainable crop growth. The undetermined fertilizer's amount is treated as a data sparsity problem that is solved primarily by adding side features such as soil fertilizer level, land size, and soil chemical properties. We specifically introduce a new method of wide matrix factorization that maps the embedded land and nutrient vectors linearly to obtain the feature representation. Furthermore, the fully connected multi-layer perceptron captures the non-linear higher-order interaction between land and nu-trients to deeply learn the historical fertilizer recommendations of the land. An exploratory outcome on two real-time datasets indicates the superior performance of the proposed method over recent state-of-the-art approaches. It reveals that the proposed deep network predicts well on nutrients data for precise fertilizer recommendation, which support farmers to increase the crop yield.