Abstract
A statistical method has been suggested for a calculation of optimal lines separating a priori obtained classes. This method is based on minimization of errors of values, which are outside their own class. The method was applied for data of V.V. Romashin and Van den Berg, and new lines dividing the classes in the optimal way were calculated. Some statistical methods have been developed for an assessment of efficiency of classifications. On the basis of these methods it has been obtained that it is statistically ineffective to distinguish a class of incomplete meandering in the Romashin's classification and it is necessary to combine it with a class of free meandering. For Van den Berg's classification it has been shown that an additional use of D 50 for a separation of single and multi-thread channels is not statistical effective too.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
