Exploring Data-Driven Approaches for Apple Disease Prediction Systems

Abstract

Apple is one of the most popular plants in the Kashmir valley of Indian origin, where it is cultivated in almost half of the horticulture area. The Kashmiri apple is well known for its deliciousness and is exported to different parts every year. However, apple plants are also susceptible to diseases such as apple scab, Alternaria leaf bloch, and apple rot. The timely detection or prediction of these diseases in the apple plants may help the farmers take appropriate measures to control the overall yield. With the emergence of artificial intelligence, machine learning-based techniques can be deployed for more accurate disease prediction. This study presents the data-driven approaches for the automated apple disease prediction system (ADPS). The study aims to explore both traditional and modern deep learning paradigms that have been used to develop ADPS for efficient and accurate disease prediction. A comparative analysis of these techniques is carried out to understand the concept deployed and datasets. Our analysis indicates that most ADPS has been designed using machine learning-based algorithms that show promising performance (i.e., 90–95%). It has been observed that Kashmiri ADPS is a least explored area where only a few studies in the literature are available. Besides, some open research issues have been identified and summarized that will serve as a reference document for the new researchers who want to explore this active field of research.