A Predictive Machine Learning and Deep Learning Approach on Agriculture Datasets for New Moringa Oleifera Varieties Prediction

Abstract

Moringa oleifera, the best known of the thirteen species of the genus Moringacae, has achieved importance due to its multipurpose usage with high nutritional value. There is very little work has been done in the advancement of moringa varieties in Pakistan.   Thus, it needs to develop a new variety of moringa with better nutritional value. The agrarian performs many experiments like interbreeding of Moringa oleifera germplasm with exotic germplasm. Furthermore, they grow it in the nursery and then move it on the field, which almost took six months in a traditional approach.  It consumes various resources and time to access the quality of newly developed varieties. This work aims to use machine learning and deep learning approaches to reduce the utilization of various resources and time which is used by the agrarian to develop a new moringa variety.  We used machine learning and deep learning approaches to make predictions about new varieties before their proper plantation.  In this research work, we took two moringa parents’ varieties with their required features like plant height, protein, potassium.  We trained machine learning and deep learning models on the feature values of parents’ varieties.  Our proposed machine learning model made the best predictions, using parents’ plant features to determine these parameter values in their offspring varieties, which will help to choose the best interbreed variety of moringa oleifera.