Evaluation of the Effectiveness of Data Preparation in Phenotyping Wheat Seedlings of Various Varieties

The work discusses the analysis of the results of recalculating the parameters of biological potential signals of plants with different ages by clustering methods of wheat seedlings “Novosibirskaya 41” varieties. Briefly described are the goals of studies of plant biopotentials, conditions and features of measurements. The results of clustering according to the attributes of biopotential signals arising from exposure to reduced and elevated temperatures on 10-, 12- and 14-day seedlings obtained using the methods of SpectralClustering, Bisecting KMeans, GaussianMixture from the scikit-learn library (formerly known as sklearn) in the Python programming environment. The need for preliminary assessment of the results of experiments is motivated for their significant difference from neighboring ones in the group in order to exclude cases of incorrect conduct of them. Partial confirmation of the hypothesis on the influence of the age of samples on the quality of their clustering by grades is recorded. Directions of further research are proposed.