МЕТОД СИНТЕЗУ ЛОГІЧНИХ ДЕРЕВ КЛАСИФІКАЦІЇ НА ПІДСТАВІ СЕЛЕКЦІЇ ЕЛЕМЕНТАРНИХ ОЗНАК

https://doi.org/10.23939/ujit2022.02.025
Надіслано: Жовтень 15, 2022
Прийнято: Жовтень 17, 2022

Ци­ту­ван­ня за ДСТУ: Пов­хан І. Ф. Ме­тод син­те­зу ло­гіч­них де­рев кла­си­фі­ка­ції на під­ста­ві се­лек­ції еле­мен­тар­них оз­нак. Ук­ра­їнсь­кий жур­нал ін­фор­ма­ційних тех­но­ло­гій. 2022, т. 4, № 2. С. 25–32.

Ci­ta­ti­on APA: Povkhan, I. F. (2022). Met­hod for synthe­si­zing lo­gi­cal clas­si­fi­ca­ti­on tre­es ba­sed on the se­lec­ti­on of ele­men­tary fe­atu­res. Uk­ra­ini­an Jo­ur­nal of In­for­ma­ti­on Techno­logy, 4(2), 25–32. https://doi.org/10.23939/ujit2022.02.025

Автори:
1
Ужгородський національний університет, м. Ужгород, Україна

Розглянута загальна задача побудови логічних дерев класифікації та розпізнавання дискретних об'єктів. Об'єктом даного дослідження є логічні дерева класифікації. Предметом дослідження є актуальні методи та алгоритми побудови логічних дерев класифікації. Метою роботи є створення простого та ефективного методу побудови моделей розпізнавання на підставі дерев класифікації для навчальних вибірок дискретної інформації, який характеризується елементарними ознаками в структурі синтезованих логічних дерев класифікації. Запропоновано загальний метод побудови логічних дерев класифікації, який для заданої початкової навчальної вибірки будує деревоподібну структуру, яка складається з набору елементарних ознак, оцінених на кожному кроці побудови моделі за даною вибіркою. Розроблено метод побудови логічного дерева, основна ідея якого полягає в апроксимації начальної вибірки довільного об'єму набором елементарних ознак. Під час формування поточної вершини логічного дерева, його вузол забезпечує виділення найбільш інформативних, якісних елементарних ознак з початкового набору. Такий підхід при побудові остаточного дерева класифікації дає змогу значно скоротити розмір та складність дерева, загальну кількість гілок та ярусів структури, підвищити якість його подальшого аналізу. Запропонований метод побудови логічного дерева класифікації дає змогу будувати деревоподібні моделі розпізнавання для широкого класу задач теорії штучного інтелекту. Розроблений та наведений в роботі метод отримав програмну реалізацію та був досліджений під час розв'язання задачі класифікації даних геологічного типу. Проведені в роботі експерименти підтвердили працездатність запропонованого математичного забезпечення та показують можливість його використання для розв'язання широкого спектру практичних задач розпізнавання та класифікації. Перспективи подальших досліджень полягають в створенні обмеженого методу логічного дерева класифікації, який полягає у введенні критерію зупинки процедури його побудови за глибиною структури, оптимізації його програмних реалізацій, а також проведення експериментального дослідження цього методу на більш широке коло практичних задач.

[1] Bodyanskiy, Y., Vynokurova, O., Setlak, G. & Pliss, I. (2015). Hybrid neuro-neo-fuzzy system and its adaptive learning algorithm, Xth Scien. and Tech. Conf. "Computer Sciences and Information Technologies" (CSIT), Lviv, 111-114.
https://doi.org/10.1109/STC-CSIT.2015.7325445
[2] Breiman, L. L., Friedman, J. H., Olshen, R. A., & Stone, C. J. (1984). Classification and regression trees. Boca Raton, Chapman and Hall/CRC, 368 p.
[3] De Mántaras, R. L. (1991). A distance-based attribute selection measure for decision tree induction. Machine learning, 6(1), 81-92.
https://doi.org/10.1023/A:1022694001379
[4] Deng, H., Runger, G., & Tuv, E. (2011). Bias of importance measures for multi-valued attributes and solutions. Proceedings of the 21st International Conference on Artificial Neural Networks (ICANN), 293-300.
https://doi.org/10.1007/978-3-642-21738-8_38
[5] Hastie, T., Tibshirani, R., & Friedman, J. (2008). The Elements of Statistical Learning. Berlin, Springer, 768 p.
https://doi.org/10.1007/978-0-387-84858-7
[6] Kamiński, B., Jakubczyk, M., & Szufel, P. (2017). A framework for sensitivity analysis of decision trees. Central European Journal of Operations Research, 26 (1), 135-159.
https://doi.org/10.1007/s10100-017-0479-6
[7] Karimi, K., & Hamilton, H. J. (2011). Generation and Interpretation of Temporal Decision Rules. International Journal of Computer Information Systems and Industrial Management Applications, 3, 314-323.
[8] Koskimaki, H., Juutilainen, I., Laurinen, P., & Roning, J. (2008). Two-level clustering approach to training data instance selection: a case study for the steel industry, Neural Networks: International Joint Conference (IJCNN-2008), Hong Kong, 1-8 June 2008: proceedings. Los Alamitos, IEEE, 3044-3049.
https://doi.org/10.1109/IJCNN.2008.4634228
[9] Kotsiantis, S. B. (2007). Supervised Machine Learning: A Review of Classification Techniques. Informatica, 31, 249-268.
[10] Laver, V. O., & Povkhan, I. F. (2019). The algorithms for constructing a logical tree of classification in pattern recognition problems. Scientific notes of the Tauride national University. Series: technical Sciences, 30(69)(4), 100-106.
https://doi.org/10.32838/2663-5941/2019.4-1/18
[11] Miyakawa, M. (1989). Criteria for selecting a variable in the construction of efficient decision trees. IEEE Transactions on Computers, 38(1), 130-141.
https://doi.org/10.1109/12.8736
[12] Painsky, A., & Rosset, S. (2017). Cross-validated variable selection in tree-based methods improves predictive performance, IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(11), 2142-2153.
https://doi.org/10.1109/TPAMI.2016.2636831
[13] Povhan, I. (2016). Designing of recognition system of discrete objects, IEEE First International Conference on Data Stream Mining & Processing (DSMP), Ukraine. Lviv, 226-231.
[14] Povhan, I. (2019). General scheme for constructing the most complex logical tree of classification in pattern recognition discrete objects . Electronics and Information Technologies, 11, 112-117.
https://doi.org/10.30970/eli.11.7
[15] Povhan, I. F. (2019). The problem of general estimation of the complexity of the maximum constructed logical classification tree. Bulletin of the national technical University Kharkiv Polytechnic Institute, 13, 104−117.
https://doi.org/10.20998/2411-0558.2019.13.10
[16] Povkhan, I. F. (2018). The problem of functional evaluation of a training sample in discrete object recognition problems. Scientific notes of the Tauride national University. Series: technical Sciences, 29(68)(6), 217-222.
[17] Povkhan, I. F. (2019). Features of synthesis of generalized features in the construction of recognition systems using the logical tree method, Materials of the international scientific and practical conference "Information technologies and computer modeling ІТКМ-2019". Ivаnо-Frаnkivsk, 169-174.
[18] Povkhan, I. F. (2019). Features random logic of the classification trees in the pattern recognition problems. Scientific notes of the Tauride national University. Series: technical Sciences, 30(69)(5), 152-161.
https://doi.org/10.32838/2663-5941/2019.5-1/22
[19] Quinlan, J. R. (1986). Induction of Decision Trees. Machine Learning, 1, 81 106.
https://doi.org/10.1007/BF00116251
[20] Srikant, R., & Agrawal, R. (1997). Mining generalized association rules. Future Generation Computer Systems, 13(2), 161 180.
https://doi.org/10.1016/S0167-739X(97)00019-8
[21] Subbotin, S. (2013). The neuro-fuzzy network synthesis and simplification on precedents in problems of diagnosis and pattern recognition. Optical Memory and Neural Networks (Information Optics), 22(2), 97-103.
https://doi.org/10.3103/S1060992X13020082
[22] Subbotin, S. A. (2013). Methods of sampling based on exhaustive and evolutionary search. Automatic Control and Computer Sciences, 47(30), 113-121.
https://doi.org/10.3103/S0146411613030073
[23] Subbotin, S. A. (2014). Methods and characteristics of localitypreserving transformations in the problems of computational intelligence. Radio Electronics, Computer Science, Control, (1), 120-128.
https://doi.org/10.15588/1607-3274-2014-1-17
[24] Subbotin, S. A. (2019). Construction of decision trees for the case of low-information features. Radio Electronics, Computer Science, Control, (1), 121-130.
https://doi.org/10.15588/1607-3274-2019-1-12
[25] Subbotin, S., & Oliinyk, A. (2017). The dimensionality reduction methods based on computational intelligence in problems of object classification and diagnosis, Recent Advances in Systems, Control and Information Technology, [eds.: R. Szewczyk, M. Kaliczyńska]. Cham, Springer, 11-19. (Advances in Intelligent Systems and Computing, 543.
https://doi.org/10.1007/978-3-319-48923-0_2
[26] Vasilenko, Y. A., Vashuk, F. G., & Povkhan, I. F. (2011). The problem of estimating the complexity of logical trees recognition and a general method for optimizing them. Scientific and technical journal "European Journal of Enterprise Technologies", 6/4(54), 24-28.
[27] Vasilenko, Y. A., Vashuk, F. G., & Povkhan, I. F. (2012). General estimation of minimization of tree logical structures. European Journal of Enterprise Technologies, 1/4(55), 29-33.
[28] Vasilenko, Y. A., Vashuk, F. G., Povkhan, I. F., Kovach, M. Y., & Nikarovich, O. D. (2004). Minimizing logical tree structures in image recognition tasks. European Journal of Enterprise Technologies, 3(9), 12-16.
[29] Vasilenko, Y. A., Vasilenko, E. Y., & Povkhan, I. F. (2002). Defining the concept of a feature in pattern recognition theory. Artificial Intelligence, 4, 512-517.
[30] Vasilenko, Y. A., Vasilenko, E. Y., & Povkhan, I. F. (2003). Branched feature selection method in mathematical modeling of multi-level image recognition systems. Artificial Intelligence, 7, 246−249.
[31] Vasilenko, Y. A., Vasilenko, E. Y., & Povkhan, I. F. (2004). Conceptual basis of image recognition systems based on the branched feature selection method. European Journal of Enterprise Technologies, 7(1), 13-15.
[32] Vtogoff, P. E. (1989). Incremental Induction of Decision Trees. Machine Learning, (4), 161−186.