Machine Learning for Varietal Binary Classification of Soybean (Glycine max (L.) Merrill) Seeds Based on Shape and Size Attributes

Creative Commons License

Çetin N.

FOOD ANALYTICAL METHODS, vol.15, pp.2260-2273, 2022 (SCI-Expanded, Scopus)

  • Publication Type: Article / Article
  • Volume: 15
  • Publication Date: 2022
  • Doi Number: 10.1007/s12161-022-02286-3
  • Journal Name: FOOD ANALYTICAL METHODS
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Agricultural & Environmental Science Database, CAB Abstracts, Compendex, Food Science & Technology Abstracts, INSPEC, Veterinary Science Database
  • Page Numbers: pp.2260-2273
  • Open Archive Collection: AVESIS Open Access Collection
  • Erciyes University Affiliated: Yes

Abstract

The most important principal quality attributes of seeds are shape, size, and mass. These parameters play a critical role in the design of classifer and grading machines. This study was conducted to develop classifcation models for distinguishing the soybean seeds based on shape, size, and mass attributes. The seeds of soybean varieties of Bravo, Ceyhan, Çevik, İlksoy, and Traksoy were classifed in pairs. Four diferent machine learning algorithms (random forest, RF; support vector machine, SVM; Naïve Bayes, NB; and multilayer perceptron, MLP) were used to evaluate the classifcation performance. In all cases, the soybean seeds of Ceyhan and Traksoy varieties were classifed with the greatest accuracy as 90.00% for the RF classifer and 89.00% for MLP. The variety pairs that followed these varieties with the highest accuracy were Çevik and İlksoy (88.00%, MLP) and Çevik and Traksoy (87.50%, RF). The highest mass (0.19 g), volume (155.02 mm3 ), geometric mean diameter (6.65 mm), and projected area (34.80 mm2 ) values were obtained from Traksoy variety. The Pillai trace and Wilks’ lambda results revealed that diferences in physical attributes of the soybean varieties were signifcant (p<0.01). In Wilks’ lambda statistics, the unexplained part of the diferences between the groups was found to be 23.0%. Traksoy and Çevik varieties with the highest Mahalanobis distances had similar attributes. Present fndings showed that MLP and RF could potentially be used for the classifcation of soybean varieties.