CASPIAN JOURNAL
MANAGEMENT AND HIGH TECHNOLOGIES
Missing data recovery methods for the issues of economic diagnostics, based on case-based reasoning
Read | Kvyatkovskaya A. E., Chertina E. V., Shurshev V. F. Missing data recovery methods for the issues of economic diagnostics, based on case-based reasoning // Caspian journal : management and high technologies. — 2018. — №4. — pp. 19-25. |
Kvyatkovskaya A. E. - Assistant, Astrakhan State Technical University, 16 Tatishchev St., Astrakhan, 414056, Russian Federation, zima00@list.ru
Chertina E. V. - Cand. Sci. (Engineering), Associate Professor, Astrakhan State Technical University, 16 Tatishchev St., Astrakhan, 414056, Russian Federation, saprikinae_1912@mail.ru
Shurshev V. F. - Doct. Sci. (Engineering), Professor, Astrakhan State Technical University, 16 Tatishchev St., Astrakhan, 414056, Russian Federation, v.shurshev@mail.ru
The study addresses the problem of the economic state valuation of an IT company, based on case-based reasoning in the context of inadequate data. The base of precedents is presented as a knowledge base of IT companies with certain characteristics and features. A proximity measure of a precedent and an analogue is used for the economic diagnostics of a peer company. A proximity measure is determined by the available characteristics in the base of precedents. The authors note that the problem of available data incompleteness can lead to structural imperfection of the data set in the base of precedents and, as a result, the impossibility or unreliability of the economic state valuation of an IT company. To solve this problem, a complex of methods for missing data processing is proposed: the deletion of objects with missing data; using of mathematical statistics tools; using of cluster analysis and classification tools. There has been developed an algorithm for valuation of analyzed peer company which tests the peer company’s characteristics with some data on it being missing. This algorithm can be used by the decision - maker for decisions making process on IT company valuation in the context of information uncertainty
Key words: IT-компания, экономическая диагностика, рассуждение по прецедентам, компания-аналог, принятие решений, алгоритмы нечеткой кластеризации, пропущенные данные, IT company, case-based reasoning, peer company, decision making, fuzzy clustering algorithms, miss