Optimizing Crop Yield Through Data-Driven Decision Making: A Comprehensive Approach to Crop Selection Using Machine Learning

Abstract

The crop cultivation generally depends on the environmental circumstances and the production of crops plays a major part to meet the increasing requirement of food. Therefore, attaining improved efficiency and sustainability is considered highly essential in farming, due to this reason, a novel Crop Recommendation System (CRS) is innovated which enables to analyse environmental and soil conditions. The introduction CRS effectively optimises the crop selection process leading to maximized crop yield, thus increasing both productivity and agricultural reliability, which is accomplished by collecting agricultural data like temperature, humidity, rainfall and soil nutrients (nitrogen, phosphorous and potassium) including pH levels. Furthermore, pre-processing is carried out in two stages which are data cleaning to prevent data inaccuracies or missing values, and label encoding which transforms categorical data into a format applicable for Machine Learning (ML) based classifier models. In addition to this, Exploratory Data Analysis (EDA) is deployed to visualize data patterns and relationships within dataset. Moreover, for model training and evaluation, the data is then split into features and labels along with train-test split. Lastly, data classification is performed using Walrus Optimized Algorithm (WaOA) based Multi-Layer Perceptron (MLP) and Bi- Directional Gated Recurrent Unit (Bi-GRU) based neural network which enables to capture sequential dependencies within the data. Therefore, the proposed system as a whole ensures improved system stability with accurate prediction and selection of crop on the basis of the environmental and soil conditions by attaining enhanced efficiency (98%), Precision (98%), Recall (98%), F1 Score (98%) respectively. Furthermore, the developed model is implemented using Python software and the attained results depicts the enhanced model performance with highly efficient classifier function.