Irkutsk, Russian Federation
A self-consistent, data-driven approach to classifying data obtained at the ISTP SB RAS mid-latitude coherent scatter radars has been developed. Based on 2021 data, a solution of the problem of automatic data classification is presented without their labeling by an expert and without postulating the number of classes. The algorithm automatically labels the data, determines the optimal number of signal classes observed by the radars, and trains a two-layer classifying neural network of an extremely simple structure. The trajectory calculations use the wave optics method and international reference models of the ionosphere and the geomagnetic field. The model is trained on signals coming from the main lobe of the antenna pattern. During training, to adapt part of the data obtained with improved spectral resolution, it is artificially coarsened to the standard resolution. Each signal class determined by the neural network is interpreted from a physical point of view, using statistical characteristics of the signals belonging to it. The number of classes in the data is demonstrated to range from 23 to 35. The significance of various parameters of the input data is assessed. It is shown that the most important parameters for the classification are the calculated scattering height and the elevation of the trajectory at the scattering point, and the least important are the spectral width of the received signal and the calculated number of reflections from the underlying surface.
decameter radar, SECIRA, ionosphere, automatic classification
1. Arnold V. On the function of three variables. American Mathematical Society Translations. 1963, pp. 51–54.
2. Barthes L., Andrè R., Cerisier J.-C., Villain J.-P. Separation of multiple echoes using a high-resolution spectral analysis for SuperDARN HF radars. Radio Sci. 1998, vol. 33, no. 4, pp. 1005–1017. DOI:https://doi.org/10.1029/98RS00714.
3. Berngardt O.I. The first comparative analysis of meteor echo and sporadic scattering identified by a self-learned neural network in EKB and MAGW ISTP SB RAS radar data. Solar-Terrestrial Physics. 2022, vol. 8, no. 4, pp. 63–72. DOI:https://doi.org/10.12737/stp-84202206.
4. Berngardt O.I. Superclustering by finding statistically significant separable groups of optimal gaussian clusters. 2023. DOI:https://doi.org/10.48550/arXiv.2309.02623.
5. Berngardt, O.I. Minimum number of neurons in fully connected layers of a given neural network (the first approximation). 2024. DOI:https://doi.org/10.48550/arXiv.2405.14147.
6. Berngardt O.I., Kutelev K.A., Potekhin A.P. SuperDARN scalar radar equations. Radio Sci. 2016, vol. 51, no. 10, pp. 1703–1724. DOI:https://doi.org/10.1002/2016RS006081.
7. Berngardt O.I., Grkovich K.V., Fedorov R.R. Synthesis of Symmetric Sounding Sequences for Ekaterinburg Coherent Decameter Radar. Radiophysics and Quantum Electronics. 2020a, vol. 62, no. 11, pp. 721–733. DOI:https://doi.org/10.1007/s11141-020-10018-y.
8. Berngardt O., Kurkin V., Kushnarev D., Grkovich K., Fedorov R., Orlov A., Harchenko V. ISTP SB RAS decameter radars. Solar-Terrestrial Physics. 2020b, vol. 6, no. 2, pp. 63–73. DOI:https://doi.org/10.12737/stp-62202006.
9. Berngardt O.I., Fedorov R.R., Ponomarenko P., Grkovich K.V. Interferometric calibration and the first elevation observations at EKB ISTP SB RAS radar at 10–12 MHz. Polar Science. 2021, vol. 28, p. 100628. DOI:https://doi.org/10.1016/j.polar.2020.100628.
10. Berngardt O.I., Kusonsky O.A., Poddelsky A.I., Oinats A.V. Self-trained artificial neural network for physical classification of ionospheric radar data. Adv. Space Res. 2022, vol. 70, no. 10, pp. 2905–2919. DOI:https://doi.org/10.1016/j.asr.2022.07.054.
11. Bilitza D., Pezzopane M., Truhlik V., Altadill D., Reinisch B.W., Pignalberi A. The International Reference Ionosphere Model: A review and description of an ionospheric benchmark. Rev. Geophys. 2022, vol. 60, no. 4, e2022RG000792. DOI:https://doi.org/10.1029/2022RG000792.
12. Blanchard G.T., Sundeen S., Baker K.B. Probabilistic identification of high-frequency radar backscatter from the ground and ionosphere based on spectral characteristics. Radio Sci. 2009, vol. 44, no. 5. DOI:https://doi.org/10.1029/2009RS004141.
13. Breiman Leo. Random forests. Machine Learning. 2001, vol. 45, no. 1, pp. 5–32. DOI:https://doi.org/10.1023/a:1010933404324.
14. Burrell A.G., Milan S.E., Perry G.W., Automatically determining the origin direction and propagation mode of high-frequency radar backscatter. Radio Sci. 2015. vol. 50, no. 12, pp. 1225–1245. DOI:https://doi.org/10.1002/2015RS005808.
15. Chisham G., Freeman M.P. A reassessment of SuperDARN meteor echoes from the upper mesosphere and lower thermosphere. J. Atmos. Solar-Terr. Phys. 2013, vol. 102, pp. 207–221. DOI:https://doi.org/10.1016/j.jastp.2013.05.018.
16. Chisham G., Lester M., Milan S.E., Freeman M.P., Bristow W.A., McWilliams K.A., et al. A decade of the Super Dual Auroral Radar Network (SuperDARN): scientific achievements, new techniques and future directions. Surveys in Geophysics. 2007, vol. 28, pp. 33–109. DOI:https://doi.org/10.1007/s10712-007-9017-8.
17. Ester M., Kriegel H.-P., Sander J., Xu X. A density-based algorithm for discovering clusters in large spatial databases with noise. V: Proc. Second International Conference on Knowledge Discovery and Data Mining. KDD96. Portland, Oregon: AAAI Press, 1996, pp. 226–231.
18. Goodfellow I., Bengio Y., Courville A. Deep learning. Adaptive Computation and Machine Learning. MIT Press. 2016.
19. Greenwald R.A., Baker K.B., Dudeney J.R., Pinnock M., Jones T.B., Thomas E.C., et al. Darn/Superdarn: A global view of the dynamics of high-latitude convection. Space Sci. Rev. 1995, vol. 71, pp. 761–796. DOI:https://doi.org/10.1007/BF00751350.
20. Greenwald R.A., Oksavik K., Barnes R., Ruohoniemi M.J., Baker J., Talaat E.R., et al. First radar measurements of ionospheric electric fields at sub-second temporal resolution. Geophys. Res. Lett. 2008, vol. 35, no. 3. DOI:https://doi.org/10.1029/2007GL032164.
21. Guo Chuan, Geoff Pleiss, Yu Sun, Weinberger K.Q. On calibration of modern neural networks. 2017. DOI: 10.48550/ arXiv.1706.04599.
22. Huang X., Kroening D., Ruan W., Sharp J., Sun Y., Thamo E., et al. A survey of safety and trustworthiness of deep neural networks: Verification, testing, adversarial attack and defence, and interpretability. Computer Sci. Rev. 2020, vol. 37, p. 100270. DOI:https://doi.org/10.1016/j.cosrev.2020.100270.
23. Hubert L., Arabie P. Comparing partitions. J. Classification. 1985, vol. 2, no. 1, pp. 193–218. DOI:https://doi.org/10.1007/bf01908075.
24. Ioffe S., Szegedy C. Batch normalization: accelerating deep network training by reducing internal covariate shift. 2015. DOI:https://doi.org/10.48550/arXiv.1502.03167.
25. Kolmogoroff A. Grundbegriffe der Wahrscheinlichkeitsrechnung. Springer Berlin Heidelberg. 1933. DOI:https://doi.org/10.1007/978-3-642-49888-6.
26. Kolmogorov A.N. On the representation of continuous functions of many variables by superposition of continuous functions of one variable and addition. Dokl. Akad. Nauk SSSR. 1957, pp. 953–956.
27. Kong Xing, Liu E., Shi S., Chen F. The implementation of deep clustering for SuperDARN backscatter echoes. Adv. Space Res. 2024, vol. 74, no. 1, pp. 243–254. DOI:https://doi.org/10.1016/j.asr.2024.03.039.
28. Kravtsov Y.A., Namazov S.A. Characteristics of scattering of radio waves from magnetically oriented inhomogeneities of the ionosphere near critical frequency. Radiotekhnika i elektronika [J. Communications Technology and Electronics]. 1980, pp. 459–466. [In Russian].
29. Kunduri B.S.R., Baker J.B.H., Ruohoniemi J.M., Thomas E.G., Shepherd S.G. An examination of SuperDARN backscatter modes using machine learning guided by ray-tracing. Space Weather. 2022, vol. 20, no. 9, e2022SW003130. DOI:https://doi.org/10.1029/2022SW 003130.
30. Lester M., Chapman P.J., Cowley S.W.H., Crooks S.J., Davies J.A., Hamadyk P., et al. Stereo CUTLASS — A new capability for the SuperDARN HF radars. Ann. Geophys. 2004, vol. 22, no. 2, pp. 459–473. DOI:https://doi.org/10.5194/angeo-22-459-2004.
31. Milan S.E., Jones T.B., Robinson T.R., Thomas E.C., Yeoman T.K. Interferometric evidence for the observation of ground backscatter originating behind the CUTLASS coherent HF radars. Ann. Geophys. 1997, vol. 15, no. 1, pp. 29–39. DOI:https://doi.org/10.1007/s00585-997-0029-y.
32. Nishitani N., Ruohoniemi J.M., Lester M., Baker J.B.H., Koustov A.V., Shepherd S.G., et al. Review of the accomplishments of mid-latitude Super Dual Auroral Radar Network (SuperDARN) HF radars. Progress in Earth and Planetary Science. 2019, vol. 6, no. 1. DOI:https://doi.org/10.1186/s40645-019-0270-5.
33. Ponomarenko P.V. Blessing Iserhienrhien i Jean-Pierre St.-Maurice. Morphology and possible origins of near-range oblique HF backscatter at high and midlatitudes. Radio Sci. 2016, vol. 51, no. 6, pp. 718–730. DOI:https://doi.org/10.1002/2016RS006088.
34. Ponomarenko P., McWilliams K.A. Climatology of HF Propagation Characteristics at Very High Latitudes From SuperDARN Observations. Radio Sci. 2023, vol. 58, no. 5, e2023RS007657. DOI:https://doi.org/10.1029/2023RS007657.
35. Ponomarenko P.V., Waters C.L., Menk F.W. Factors determining spectral width of HF echoes from high latitudes. Ann. Geophys. 2007, vol. 25, no. 3, pp. 675–687. DOI:https://doi.org/10.5194/angeo-25-675-2007.
36. Ponomarenko P.V., Koustov A.V., St.-Maurice J.-P., Wiid J. Monitoring the F-region peak electron density using HF backscatter interferometry. Geophys. Res. Lett. 2011, vol. 38, no. 21. DOI:https://doi.org/10.1029/2011GL049675.
37. Ribeiro A.J., Ruohoniemi J.M., Baker J.B.H., Clausen L.B.N., de Larquier S., Greenwald R.A. A new approach for identifying ionospheric backscatter in midlatitude SuperDARN HF radar observations. Radio Sci. 2011, vol. 46, no. 4. DOI:https://doi.org/10.1029/2011RS004676.
38. Ribeiro A.J., Ruohoniemi J.M., Ponomarenko P.V., Clausen L.B.N., Baker J.B.H., Greenwald R.A., Oksavik K., de Larquier S. A comparison of SuperDARN ACF fitting methods. Radio Sci. 2013, vol. 48, no. 3, pp. 274–282. DOI:https://doi.org/10.1002/rds.20031.
39. Rousseeuw P.J. Silhouettes: A graphical aid to the interpretation and validation of cluster analysis. J. Computational and Applied Mathematics. 1987, vol. 20, pp. 53–65. DOI:https://doi.org/10.1016/0377-0427(87)90125-7.
40. Rumelhart D.E., Hinton G.E., Williams R.J. Learning Internal Representations by Error Propagation. V: Parallel Distributed Processing, Volume 1: Explorations in the Microstructure of Cognition: Foundations. The MIT Press. 1986. DOI:https://doi.org/10.7551/mitpress/5236.003.0012.
41. Saxena Amit, Mukesh Prasad, Akshansh Gupta, Neha Bharill, Om Prakash Patel, Tiwari A., et al. A review of clustering techniques and developments. Neurocomputing. 2017, vol. 267, pp. 664–681. DOI:https://doi.org/10.1016/j.neucom.2017.06.053.
42. Schwarz Gideon. Estimating the dimension of a model. The Annals of Statistics. 1978, vol. 6, no. 2, pp. 461–464. DOI:https://doi.org/10.1214/aos/1176344136.
43. Shorten C., Khoshgoftaar T.M. A survey on image data augmentation for deep learning. J. Big Data. 2019, vol. 6, no. 1. DOI:https://doi.org/10.1186/s40537-019-0197-0.
44. Smiti A., Elouedi Z. Fuzzy density based clustering method: Soft DBSCAN-GM. V: 2016 IEEE 8th International Conference on Intelligent Systems, 2016, pp. 443–448. DOI:https://doi.org/10.1109/IS.2016.7737459.
45. Sutton R.S., Barto A.G. Reinforcement Learning: An Introduction. Second. The MIT Press. 2018.
46. Vallés-Pérez I., Soria-Olivas E., Martıénez-Sober M., Serrano-López A.J., Vila-Francés J., Gómez-Sanchıés J. Empirical study of the modulus as activation function in computer vision applications. Engineering Applications of Artificial Intelligence. 2023, vol. 20, p. 105863. DOI:https://doi.org/10.1016/j.engappai.2023.105863.
47. URL: https://github.com/berng/WrappedClassifier/tree/master/v.3.0 (accessed April 15, 2025).
48. URL: http://sdrus.iszf.irk.ru/node/107 (accessed April 15, 2025).
49. URL: https://rscf.ru/project/24-22-00436/ (accessed April 15, 2025).
