Machine Learning Application for Classification of Rice Varieties: A Systematic Literature Review

The diversity of rice creates challenges in identifying rice varieties. Rice quality is usually judged by price, which does not always reflect the actual quality. This research aims to identify the most optimal and efficient algorithm for rice classification. It can also highlight research gaps that can be used to develop further research questions. The approach to searching and selecting articles followed the PRISMA guidelines (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) are a set of evidence-based recommendations designed to improve the quality and transparency of reporting in systematic reviews and meta-analyses. The research was conducted through a systematic literature review that focused on factors important for identifying rice types based on shape. The process was divided into several steps: selecting the database, reviewing the objectives and research questions, creating a search strategy, establishing search criteria, defining inclusion and exclusion criteria, and extracting data. The results were examined and analyzed to answer the research questions. This research involved the classification of rice types using QNN (Quantized Neural Network). Quantized Neural Networks (QNN) offer high computational efficiency, making them ideal for edge devices, but this technology is still in the early stages of development and requires specialized hardware.