PECULIARITIES OF USING MACHINE LEARNING TOOLS DURING IDENTIFICATION OF SIMILAR PLANT VARIETIES (Lactuca sativa L. var. capitata as example)

Abstract

Technological approaches to using Machine Learning tools in the identification of new crop varieties based on the data set of well-known varieties, by using the algorithm of the nearest neighbor were consid-ered on the example of morphological description data of Lactuca sativa L. var. capitata varieties. The aim of the work is to give an example of Machine Learning method and assess the suitability of its using while processing data on codes of plant varieties’ morphological description. To substantiate the suitability of us-ing machine learning in processing the results of examining difference, uniformity and stability, in particular the identification of groups of similar varieties of plants analytical and statistical methods were used. The information technology of using machine learning tools for the formation of a computer model of similar lettuce varieties by applying the IBM SPSS Statistics statistical package has been developed. The results of the experiment with a computer-based training model revealed that the most accurate classification results were obtained by using the sign of the size of lettuce head as the target variable model and the sign of lettuce head density as the focal variable. The study showed the suitability of using the Machine Learning tool to identify groups of similar lettuce varieties by morphological features. The IBM SPSS Statistics package of statistical programs is convenient in using, it provides the researcher with a wide range of tools for experi-menting with a botanical taxon variety model, and enables to visualize the obtained modeling results using diagrams that clearly show the results. The main diagram of the model is interactive, allowing the research-er to experiment with the model. This method can be recommended for using in processing qualification ex-pertise data for differences, uniformity and stability.