Cocoa Beans Quality Prediction Using Near-Infrared Spectroscopy and Several Machine Learning Techniques

This research paper presents an innovative approach to cocoa beans quality testing using near-infrared (NIR) spectra and deep learning techniques. The production of high-quality chocolate products depends on the quality of cocoa beans, which are a vital element in the business. And the quality of cocoa beans plays a critical role in the production of high-quality chocolate products. Traditional methods of cocoa beans quality testing are time-consuming and labor-intensive, which can result in increased costs and reduced efficiency. By precisely predicting the quality of cocoa beans using NIR spectra and deep learning algorithms, the suggested method seeks to get over these restrictions. This study uses NIR spectra data from samples of various quality cocoa beans, and several machine learning techniques (like Linear Regression, SVR, Ridge Regression, and SGDR) are trained using the data. The model is then put to the test on a different set of data to determine how well it performs. For moisture, SVR performed best with 0.8125 R2 score whereas for fat content Ridge Regression outperformed others with R2 score of 0.6910. The results show that the suggested technique can reliably predict the quality of cocoa beans, which can help improve the efficiency and cost-effectiveness of the chocolate industry. The results of this study holds significant implications for the cocoa beans sector, as well as holds the potential to raise the quality of chocolate goods.