Design of Trend-Identification Algorithms with Learning (Algebraic Approach)

Section 1 presents the main questions associated with the application of the algebraic approach to pattern recognition and classification. The answers to these questions constitute a problem-oriented theory of this application domain. In Sections 2 and 3, the listed questions are used to construct a solvability theory for the design of trend-identification algorithms. We formally describe the initial information set (algorithm inputs), the final information set (algorithm outputs), the precedents, and the additional constraints. Solvability and regularity criteria are derived for the problem of construction of trend-identification algorithms.