No abstract

No abstract

In the article the ideas, progress, and results of a theoretic-and-applied research direction are described, aimed at highly formalized method creation for collinearity analysis, devoted to functional and statistical relationships co-assessment that belong to socio-economic and other human activities areas. With respect to related mathematical-statistical methods of correlation and regression analyses the new method differs by some principal novelties implemented there: (1) opportunity of dependencies being compared mutual adaptation in space of attributes that include several parameters being optimized -dependencies scales ratio, their mutual shift, methods of interpolation, extrapolation, normalization, and censoring with respect to dependencies numeric values and their samples; (2) variable Minkowsky metric, used for regression errors or functional residuals summation mechanism's management; (3) utilization of effective combined algorithm for solution attaining, which consists in finding the target function's global minimum, based on combination of Monte-Carlo and steepest descent methods. Results of the method and its program realization variants' statistical tests are briefly characterized, preferable directions of the method's and its inner formalized procedures' further improvement are pointed out. КЛЮЧЕВЫЕ СЛОВА Анализ коллинеарности, регрессия, взаимная корреляция, метод Монте-Карло.