|
Library
|
Your profile |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.|
Software systems and computational methods
Reference:
Bakhrushin V.E.
Software implementation of non-linear statistics relationships analysis methods in the
R system
// Software systems and computational methods.
2014. ¹ 2.
P. 228-238.
URL: https://en.nbpublish.com/library_read_article.php?id=65265
Bakhrushin V.E. Software implementation of non-linear statistics relationships analysis methods in the R systemAbstract: existing software for data statistical analysis (SPSS, Statistica etc.) usually offer for defining correlations just methods applicable for finding linear relationships in numerical data, along with some relation indicators for rank, qualitative and mixed data. However actual relation between quantitative data is often nonlinear. This leads to the fact that present means do not allow identifying such relations, which can lead to false conclusions about the absence of correlation. An universal indicator of present statistical correlation between two rows of numerical data is sample coefficient of determination. There are two approaches to calculated that coefficient: first is based on the approximation of some unknown function with piecewise constant function, second is based on the smoothing available data. The article proposes software realization for both methods in R system. The advantage of this system is in the availability of a large number of specialized library functions for statistical analysis, as well as in writing programs for non-standard tasks. Testing of the developed application on model examples proved their correctness allowing the use for solving practical problems in nonlinear correlation analysis. Keywords: nonlinear relationship, coefficient of determination, software, R programming Language, data smoothing, correlation ratio, Pearson correlation coefficient, testing, data grouping, piecewise constant function
This article can be downloaded freely in PDF format for reading. Download article
References
1. Bakhrushin V.ª. Metodi anal³zu danikh. – Zapor³zhzhya: KPU, 2011. – 268 s.
2. Gaydyshev I. Analiz i obrabotka dannykh. Spetsial'nyy spravochnik – SPb.: Piter, 2001. – 752 s. 3. Byuyul' A., Tsefel' P. SPSS: Iskusstvo obrabotki informatsii. Analiz statisticheskikh dannykh i vosstanovlenie skrytykh zakonomernostey. – SPb.: DiaSoftYuP, 2005 – 608 s. 4. Khalafyan A.A. Statistica 6. Statisticheskiy analiz dannykh. – M.: OOO Binom-Press, 2008. – 512 s. 5. Kendel M. Rangovye korrelyatsii. – M.: Statistika, 1975. – 216 s. 6. Gauthier T. D. Detecting Trends Using Spearman’s Rank Correlation Coefficient // Environmental Forensics. – 2001. No 2. – P. 359 – 362. 7. Bakhrushin A.V., Bakhrushin V.E. Testirovanie gipotez o nelineynykh svyazyakh s ispol'zovaniem yazyka programmirovaniya R // Sistemnye tekhnologii: Reg³onal'niy m³zhvuz³vs'kiy zb³rnik naukovikh prats'. Dn³propetrovs'k, 2013. – ¹ 3(86). – S. 168 – 172. 8. Bakhrushin V.E. Metody otsenivaniya kharakteristik nelineynykh statisticheskikh svyazey // Sistemn³ tekhnolog³¿: Reg³onal'niy m³zhvuz³vs'kiy zb³rnik naukovikh prats'. Dn³propetrovs'k, 2011. – ¹ 2(73). – S. 9 – 14. 9. Anderson T. Statisticheskiy analiz vremennykh ryadov. – M.: Mir, 1976. – 755 s. 10. Bakhrushin V.ª., Pavlenko V.ª., Petrova S.V. Zastosuvannya pokaznik³v nel³n³yno¿ korelyats³¿ dlya pobudovi y anal³zu kros-korelyats³ynikh funkts³y // Skladn³ sistemi ³ protsesi. – 2009, ¹ 2. – S. 78 – 85. 11. Bakhrushin V.E., Pavlenko V.E., Petrova S.V. Primenenie vyborochnogo koeffitsienta determinatsii dlya postroeniya i analiza kross-korrelyatsionnykh funktsiy // Fundamental'nye fiziko-matematicheskie problemy i modelirovanie tekhniko-tekhnologicheskikh sistem / Pod red. Yu.M. Solomentseva, B.N. Chetverushkina, A.V. Bogolyubova i dr. – M.: MGTU “STANKIN”, Yanus-K, 2010. – Vyp. 13. – S. 4 12. Statisticheskiy analiz dannykh v sisteme R / A.G. Bukhovets, P.V. Moskalev, V.P. Bogatova, T.Ya. Biryuchinskaya; Pod red. prof. Bukhovtsa A.G. – Voronezh: VGAU, 2010. – 124 s |
© 1998 – 2024 Nota Bene. Publishing Technologies. NB-Media Ltd.
