2025, Vol. 10, Special Issue 1
Exploring and comparing outlier detection approaches for apple data
Author(s): Shahran Dahlan, SA Mir, TA Raja, Nageena Nazir, MS Pukhta, SA Simnani and K Gautam
Abstract: The effectiveness of four outlier detection algorithms—Elliptical Envelope, Isolation Forest, One-Class SVM, and Local Outlier Factor (LOF)—on Gala Red Lum apple variety, focusing on yield and trunk circumference area (TCA) was studied. The analysis was conducted in both univariate and bivariate phases, utilizing key statistical metrics. Outlier removal significantly improved data symmetry and reduced variability, enhancing regression accuracy. The bi-variate analysis performed better than uni-variate analysis as it yielded higher R² values and lower error metrics, offering valuable insights for data analysis. Results indicate that the Elliptical Envelope algorithm (R2 =0.93) outperformed the others, while the One-Class SVM showed limitations.
Pages: 36-40 | Views: 50 | Downloads: 4Download Full Article: Click HereHow to cite this article:
Shahran Dahlan, SA Mir, TA Raja, Nageena Nazir, MS Pukhta, SA Simnani, K Gautam. Exploring and comparing outlier detection approaches for apple data. Int J Stat Appl Math 2025;10(1S):36-40.