INFORMATION AND TECHNICAL FEATURES OF THE TEST FOR DISTINCTNESS OF NEW VARIETIES OF LACTUCA SATIVA L. VAR. CAPITATA

Purpose. To substantiate the information and technical features of the use of IBM SPSS Statistics tools in determ­ining the distinctiveness criterion for new varieties of lettuce Lactuca sativa L. var. capitata based on the morphological code formulas of phenotypes of well-known varieties using the nearest neighbor algorithm in a group of similar varieties.

Methods. Analytical, which is based on a comparison of the methods and means of data mining obtained by identification – a morphological description of the variety with the subsequent use of descriptive and multidimensional statistics. In accordance with the «Method for the examination of varieties of lettuce Lactuca sativa L. for distinctness, uniformity and stability» signs that do not vary or very poorly vary were used to group varieties. These signs are used seperately or in combination with others.

Results. As a result of modeling using the SPPS package, several mo­dels of similar varieties of Lactuca sativa L. var. сapitata were formed. The total number of varieties in the sample is distributed as follows: 71.4% represented the training sample, and 28.6% represented the control one. The resulting data is visualized on the diagrams of the model of the largest similarity. It is necessary to take into account the type of expression of the studied characteristic (sign): in a qualitative, quantitative, pseudo-qualitative way. A simulation experiment with a model of similar varieties of head-lettuce showed the dependence of the result on the selected target variable. The target variables were signs «lant: head formation», «head by density», «seed: coloration», «head size».

Conclusions. The effectiveness of technological tools for analyzing the examination data on distinctness, uniformity and stabili­ty (DUS) has been revealed, which greatly facilitates the search for patterns among a large data set. Differentiation of data into training and control allows you to «train» the model on a data set of well-known varieties. When building a model, it is important to correctly determine the target and focal variables. IBM SPSS Statistics software package is recommended as a tool.