| Issue № 1 |
Methods of ecological investigations |
pdf-version |
Introduction to probabilistic neural networks for ecologists
| Korosov Andrey Victorovich | DSc, professor, Petrozavodsk State University, 33 Lenin Ave., Petrozavodsk, Republic of Karelia, 185910, Russia, korosov@psu.karelia.ru |
| Marfitsyna Natalya Alexandrovna | Petrozavodsk State University, 33 Lenin Ave., Petrozavodsk, Republic of Karelia, 185910, Russia, marfitsyna.nata@mail.ru |
|
Keywords: neural networks euclidean distance gaussian distribution density sexual dimorphism image interpretation |
Summary: The article examines the algorithm for constructing and calculating a probabilistic neural network (PNN)using examples from ecology. First, the components and key concepts (distance, kernel, neuron) are considered separately, followed by the key stages of the entire technology. The network structure is presented in more detail: in addition to the typical four layers (input, radial, summation, output), the neuron layer is represented by three sub-layers. This explains the variety of terms used to denote it in various publications. The concept of a radial basic activation function is examined in detail using examples. The role of a single model parameter, the kernel diameter, is discussed. As an example, two ecological problems are solved: the sexual identification of animals and the decoding of different animal habitats from a satellite image. In addition to the main description, a solution to the problem using AI technologies is presented. A technology for assessing the effectiveness of classification and ways to optimize these solutions are considered. The application of the PNN package in the R environment for constructing probabilistic neural networks is demonstrated. © Petrozavodsk State University |
Received on: 27 December 2025 Published on: 27 March 2026 | |
References
AI. Alice AI for solving real problems. 2025. URL: https://alice.yandex.ru/?utm_campaign=ntp_new_chat_btn&utm_source=desktop_browser (data obrascheniya: 20.12.2025).
Bandwidth Selectors for Kernel Density Estimation, R Documentation. 2025. URL: https://stat.ethz.ch/R-manual/R-devel/library/stats/html/bandwidth.html (data obrascheniya: 24.11.2025).
Bogdanova A. Activation functions – what are they and why the neural network doesn't work without them, DTF. 2025. URL: https://dtf.ru/id2687299/4076435-funktsii-aktivatsii-v-nevrosnetyakh (data obrascheniya: 24.11.2025).
Borovikov V. P. Neural networks. STATISTICA Neural Networks: Methodology and technologies of modern data analysis. M.: Goryachaya liniya - Telekom, 2008. 392 p. URL: https://z-library.la/book/2425255/574e37/Neyronnye-seti-statistica-neural-networks-Metodologiya-i-tehnologii-sovremennogo-analiza-dannyh.html?dsource=recommend (data obrascheniya: 24.11.2025).
Caregorodcev V. G. Optimization of data preprocessing: Lipschitz constant of training samples and properties of trained neural networks, Neyrokomp'yutery: razrabotka, primenenie. 2003. No.7. P. 3–8. URL: http://neuropro.ru/mypapers/neurcompmag03_1.pdf (data obrascheniya: 24.11.2025).
Dolenko S. A. BF. Neural networks based on RBF. Probabilistic networks and networks with general regression. Kohonen networks and self-organizing maps, Mashinnoe obuchenie. Iskusstvennye neyronnye seti i geneticheskie algoritmy. Teach-in. 2015. URL: https://teach-in.ru/lecture/2023-03-10-Dolenko (data obrascheniya: 15.10.2025).
Guseva T. L. Korosov A. V. Bespyatova L. A. Anikanova V. S. Long-term dynamics of the biotopic distribution of the common shrew (Sorex araneus, Linnaeus 1758) in mosaic landscapes of Karelia, Uchenye zapiski Petrozavodskogo gosudarstvennogo universiteta. 2014. No. 8 (145). T. 2. P. 13–20. URL: https://sciup.org/uchzap-petrsu/2014-8-145-2 (data obrascheniya: 24.11.2025).
Hagan M. T., Demuth H. B., Beale M.: Neural Network Design. Beijing: PWS Publishing Company, 2002. 1012 p. URL: https://github.com/thanhsmind/machine-learning-books/blob/master/Neural%20Network%20Design%202nd%20edition%202014.pdf (data obrascheniya: 24.11.2025).
Hajmeer M., Basheer I. A probabilistic neural network approach for modeling and classification of bacterial growth/no-growth data, Journal of Microbiological Methods. 2002. Vol. 51, Issue 2. P. 217–226. DOI: 10.1016/S0167-7012(02)00080-5
Introduction to system analysis: application in ecology. M.: Mir, 1981. 256 p. URL: https://vk.com/wall-184903207_6650?ysclid=mj7pa6901y103281711 (data obrascheniya: 15.10.2025).
Kallan R. Basic concepts of neural networks. M.: Izd. dom «Vil'yams», 2001. 288 p. URL: https://vk.com/wall-6509366_1080 (data obrascheniya: 15.10.2025).
Kernel density estimation, Wikipedia. URL: https://en.wikipedia.org/wiki/Kernel_density_estimation#Bandwidth_selection (data obrascheniya: 24.11.2025).
Korosov A. V. Bugmyrin S. V. Burdova T. L. Kireeva M. L. Lapina S. A. An information system for studying vertebrates: the experience of creation and use, Trudy KarNC RAN. No 8. Ser. Ekologicheskie issledovaniya. 2022. C. 123–133. DOI: 10.17076/eco1614
Korosov A. V. Marficyna N. A. Decryption of animal habitats using deep learning methods of the KERAS library. The InterCarto. InterGIS. M.: Geograficheskiy fakul'tet MGU, 2025. T. 31, ch. 2. P. 54–65. DOI: 10.35595/2414-9179-2025-2-31-54-65
Korosov A. V. Ecology of the common viper (Vipera berus L.) in the North (facts and models). Petrozavodsk: Izd-vo PetrGU, 2010. 264 p.
Korosov A. V. Neural networks for ecology: an introduction, Principy ekologii. 2023. No. 3. P. 76–96. DOI: 10.15393/j1.art.2023.14002
Korosov A. V. On the contradiction between statistical parameters of population dynamics, Principy ekologii. 2012. No. 2. P. 53–58. DOI: 10.15393/j1.art.2012.1141
Korosov A. V. The meaning and applicability of nuclear methods in environmental research, Principy ekologii. 2024. No. 4. P. 59–90. DOI: 10.15393/j1.art.2024.15662
Merfi K. P. Probabilistic machine learning: an introduction. M.: DMK Press, 2022. 940 p. URL: https://vk.com/wall-1172233_62640 (data obrascheniya: 24.11.2025).
Munoz-Mas R., Fukuda S., Portoles J., Martinez-Capel F. Revisiting probabilistic neural networks: a comparative study with support vector machines and the microhabitat suitability for the Eastern Iberian chub (Squalius valentinus), Ecological Informatics. 2018. Vol. 43. P. 24–37. DOI: 10.1016/j.ecoinf.2017.10.008
Nazin P. S. Gotovcev P. M. Using probabilistic neural networks to predict protein localization in cellular compartments, Matematicheskaya biologiya i bioinformatika. 2019. T. 14, No. 1. P. 220–232. DOI: 10.17537/2019.14.220
Neural network: unlocking the possibilities of artificial intelligence, Easiio. 2025. URL: https://ru.easiio.com/probabilistic-neural-network/ (data obrascheniya: 24.11.2025).
Package pnn. Probabilistic neural networks, R Documentation. 2013. P. 1–9. URL: https://r2013-lyon.sciencesconf.org/file/pnn.pdf (data obrascheniya: 05.10.2025).
Robertson S. G., Morison A. K. Age Estimation of Fish Using a Probabilistic Neural Network, Ecological Informatics. Understanding Ecology by Biologically-Inspired Computation. Springer, 2003. P. 369–382. DOI: 10.1007/978-3-662-05150-4_19 (data obrascheniya: 24.11.2025).
Rostovcev V. S. Artificial neural networks. Kirov: Izd-vo VyatGU, 2014. 208 p. URL: http://iweb.vyatsu.ru/document/material/41/_Uchebnik%20INS_2014_E4743.pdf (data obrascheniya: 24.11.2025).
Sanzhapov B. H. Application of a probabilistic neural network for express analysis of the ecological state of the atmosphere of an urban roadside area, Inzhenerno-stroitel'nyy vestnik Prikaspiya: Nauchnotehnicheskiy zhurnal. 2025. No. 1 (51). P. 127–131. DOI: 10.52684/2312-3702-2025-51-1-127-131
Shelomenceva I. G. Classification of microscopic sputum images using probabilistic Bayesian neural networks, Ekonomika. Informatika. 2022. T. 49, No. 3. P. 575–581. DOI: 10.52575/2687-0932-2022-49-3-575-581
Shitikov V. K. Mastickiy S. E. Classification, regression, and other Data Mining algorithms using R. Tol'yatti; London, 2017. 351 p. URL: https://github.com/ranalytics/data-mining (data obrascheniya: 24.11.2025).
Shitikov V. K. Rozenberg G. S. Randomization and bootstrap: statistical analysis in biology and ecology using R. Tol'yatti: Kassandra, 2013. 314 p. URL: https://en.pdfdrive.to/book/r-78 (data obrascheniya: 24.11.2025).
Sholle F. Deep learning on R. SPb.: Piter, 2018. 400 p.
Silverman B. W. Density Estimation for Statistics and Data Analysis, Biometrical Journal. 1986. Vol. 30, Issue 7. 22 p. URL: https://en.wikipedia.org/wiki/Kernel_density_estimation#cite_note-SI1998-22 (data obrascheniya: 24.11.2025).
Specht D. F. Probabilistic Neural Networks, Neural Networks. 1990. Vol. 3. P. 109–118. DOI: 10.1016/0893-6080(90)90049-Q. URL: https://wiki.eecs.yorku.ca/course_archive/2010-11/W/4403/_media/specht1990pnn.pdf (data obrascheniya: 24.11.2025).
Teles L. O., Fernandes M., Amorim J., Vasconcelos V. Video-tracking of zebrafish (Danio rerio) as a biological early warning system using two distinct artificial neural networks: Probabilistic neural network (PNN) and self-organizing map (SOM), Aquatic Toxicology. 2015. Vol. 165. P. 241–248. DOI: 10.1016/j.aquatox.2015.06.008
Terry A. M. R., McGregor P. K. Census and monitoring based on individually identifiable vocalizations: the role of neural networks, Animal Conservation. 2002. Vol. 5. P. 103–111. DOI: 10.1017/S1367943002002147
USGS science for a changing world, EarthExplorer. 2023. URL: https://earthexplorer.usgs.gov/ (data obrascheniya: 20.11.2025).
Veselov O. V. Fuzzy logic and neural networks in control and diagnostic systems. Vladimir: Izd-vo Vladim. gop. un-ta, 2023. 288 p. URL: https://dspace.www1.vlsu.ru/handle/123456789/10985 (data obrascheniya: 11.10.2025).
Zvyagin M. Yu. Bespalov M. S. Aleksandrov A. V. Applied algorithms on graphs. Vladimir: Izd-vo Vladim. gop. un-ta, 2005. 44 p. URL: https://dspace.www1.vlsu.ru/handle/123456789/462 (data obrascheniya: 15.10.2025).
© 2011 - 2026
Petrozavodsk State University
technical support - RCNIT
The editor-in-chief - Andrey Korosov
ecopri@petrsu.ru