An integrated KNN-GIS framework for high-accuracy spatial prediction of shallow foundation bearing capacity in complex geological terrains: a case study from Pokhara Valley, Nepal

Abstract

Accurate spatial prediction of shallow foundation bearing capacity remains challenging in geologically complex terrains where conventional interpolation methods fail to adequately capture abrupt geological transitions. This study introduces an integrated KNN-GIS framework that combines predictive modeling with quantitative validation and systematic bias correction. Using data of 410 boreholes in the Pokhara Valley, Nepal, an optimized KNN model was developed and validated through comprehensive stratum-wise geological analysis. The framework establishes a validation system that categorizes geological formations into four classes (A-D) based on predictive accuracy, identifies systematic biases in karst-prone formations, and provides mathematically derived correction factors to address these limitations. A three-phase engineering decision system translates validation results into actionable investigation protocols, producing spatially explicit bearing capacity maps with quantified uncertainty bounds. The methodology successfully delineates high-risk karst zones from stable bedrock areas while offering a clear and reproducible workflow for preliminary site assessment. This integrated approach bridges machine learning predictions with engineering reliability, providing a robust, transferable tool for risk-informed urban development in geologically challenging environments worldwide.

Keywords

Machine Learning / K-Nearest Neighbor / Geographic Information System / Spatial Prediction / Bearing capacity

References

[1] Shakir RR. Undrained Bearing Capacity For Strip Footing On Two Layered Elastic-Plastic Clay By Using FEM. Eng Technol J, 2005, 24(1): 651-669

[2] . McGregor JA, Duncan JM (1998) Performance and use of the standard penetration test in geotechnical engineering practice. Virginia: Virginia Polytechnic Institute and State University

[3] Civelekler E. Using GIS for the allowable soil bearing capacity estimation according to the Terzaghi (1943) equation in Eskişehir city center, Türkiye. Int J Eng Geosci, 2023, 8(3): 310-317

[4] Ali HM, Shakir RR. Geotechnical map of Thi Qar governorate using geographical information systems (GIS). Mater Today Proc, 2022, 60: 1286-1296

[5] Tomczak M. Spatial interpolation and its uncertainty using automated anisotropic inverse distance weighting (IDW)-cross-validation/jackknife approach. J Geogr Inf Decis Anal, 1998, 2(2): 18-30

[6] Pishbin M, Fathianpour N, Mokhtari AR. Uniaxial compressive strength spatial estimation using different interpolation techiques. Int J Rock Mech Min Sci, 2016, 89: 136-150

[7] Karki BS, Chao KC, Manandhar S, Wong RK (2023) Prediction of undrained shear strength in soil-cement columns constructed with dry mixing method for Bangkok soft clay. In Smart Geotechnics for Smart Societies (pp.696-701). Astana: CRC Press

[8] Fan S, Zhang Y, Li S, Han M (2026) A machine learning based framework for the probability failure envelope analysis of foundations in spatially variable clay. Engineering Applications of Artificial Intelligence. 143:110005. https://doi.org/10.1016/j.engappai.2025.110005

[9] Elsawy M, Alsharekh M, Shaban M. Modeling undrained shear strength of sensitive alluvial soft clay using machine learning approach. Appl Sci, 2022, 12(19): 1-11

[10] Huang T, Wan C, Liu T, Miao C. A data-driven prediction method for shear bearing capacity of flexural-shear type RC column based on PCC-XGB-SHAP. Case Stud Constr Mater, 2025

[11] Suppakul R, Sangjinda K, Jitchaijaroen W, Phuksuksakul N, Keawsawasvong S, Nuaklong P (2025) Bearing capacity prediction of open caissons in two-layered clays using five tree-based machine learning algorithms. Intelligent Geoengineering. p 11

[12] Gladious J, Paul PS, Mukhopadhyay M. Machine learning based prediction of geotechnical parameters affecting slope stability in open-pit iron ore mines in high precipitation zone. Sci Rep, 2025

[13] Meyerhof GG. Influence of roughness of base and ground water conditions on the ultimate bearing capacity of foundations. Géotechnique, 1955, 5(3): 227-242

[14] Nwaila GT, Zhang SE, Bourdeau JE, Frimmel HE, Ghorbani Y. Spatial interpolation using machine learning: from patterns and regularities to block models. Nat Resour Res, 2024, 33(1): 129-161

[15] Ver Hoef JM, Temesgen H. A comparison of the spatial linear model to nearest neighbor (k-NN) methods for forestry applications. PLoS ONE, 2013, 8(3): e59129

[16] Ahmed MS, N’diaye M, Attouch MK, Niange SD. K-nearest neighbors prediction and classifification for spatial data. J Spatial Econ, 2023, 4112

[17] Zhang M, Shi W, Xu Z. Systematic comparison of five machine-learning models in classification and interpolation of soil particle size fractions using different transformed data. Hydrol Earth Syst Sci, 2020, 2452505-2526

[18] Verma G, Kumar B, Kumar C, Ray A, Khandelwal M. Application of KRR, K-NN and GPR algorithms for predicting the soaked CBR of fine-grained plastic soils. Arab J Sci Eng, 2023, 48(10): 13901-13927

[19] Suleymanov A, Tuktarova I, Belan L, Suleymanov R, Gabbasova I, Araslanova L. Spatial prediction of soil properties using random forest, k-nearest neighbors and cubist approaches in the foothills of the Ural Mountains, Russia. Model Earth Syst Environ, 2023, 9(3): 3461-3471

[20] Cao N, Yan XE, Zhang L, Xu G, Ma J. Hybrid k-nearest neighbors models with metaheuristic optimization for predicting undrained shear strength. Informatica, 2025, 49(25): 20

[21] Skempton AW. Standard penetration test procedures and the effects in sands of overbudden pressure, relative density, particle size, aegin and overconsolidation. Geotechnique, 1986, 363425-447

[22] Liao SS, Whitman RV. Overburden correction factors for SPT in sand. J Geotech Eng, 1986, 112(3): 373-377

[23] Peck RB, Hanson WE, Thornburn TH (1991) Foundation Enginnering. Canada: John Wiley & Sons

[24] Bowles JE, Guo Y. Foundation Analysis and Design, 19965New York, McGraw-hill567

[25] Khan F, Das B, Mishra SRK, Awasthy M. A review on the feasibility and application of geospatial techniques in geotechnical engineering field. Mater Today Proc, 2022, 49: 311-319

[26] Setianto A, Triandini T (2013) Comparision of Kriging and Inverse Distance Weighted (IDW) Interpolation Methods In Lineament Extraction and Analysis. J Appl Geol 5(1):21–29

[27] Thapa D, Paudyal KR, Sah RB, Dhungana U, Chetri RK. A revised Quaternary stratigraphy in and around Pokhara valley, Gandaki province, Nepal. Himalayan Geol, 2023, 44(2): 89-103

[28] Yamanaka H, Yoshida M, Arrita K. Terrace landform and quaternary deposit around Pokhara Valley, Central Nepal. J Nepal Geol Soc, 1982, 2: 113-142

[29] DMG (1998) Geology and soil map of Pokhara Metropolitan City. Department of Mines and Geology, Government of Nepal, Kathmandu

[30] I 2131 (2002) Method for Standard Penetration Test for Soils, New Delhi: Bureau of Indian Standards

[31] NBC:205 (2024) Ready-to-use detailing guideline for low rise reinforced concrete buildings without masonry infill, Kathmandu: Department of Urban Development and Building Construction, Ministry of Urban Development, 2024.

[32] Meyerhof GG. Some recent research on the bearing capacity of foundations. Can Geotech J, 1963, 1(1): 16-26

[33] Maleika W. Inverse distance weighting method optimization in the process of digital terrain model creation based on data collected from a multibeam echosounder. Appl Geomat, 2020, 12(4): 397-407

[34] Nusret D, Dug S (2012) Applying the inverse distance weighting and kriging methods of the spatial interpolation on the mapping the annual precipitation in Bosnia and Herzegovina. International Congress on Environmental Modelling and Software

[35] Deutsch C, Journel A. GSLIB: Geostatistical Software Library and User's Guide, 19981New York, Oxford University Press83108

[36] Galupino J, Dungca J (2022) Development of a k-Nearest Neighbor (kNN) Learning Model to Estimate the SPT N-Values of Valenzuela City, Philippines. In 9th AUN/SEED- Net Regional Conference on Natural Disaster (RCND), Kuala Lumpur

[37] Shdefat AY, Mostafa N, Al-Arnaout Z, Kotb Y, Alabed S. Optimizing HAR systems: comparative analysis of enhanced SVM and k-NN classifiers. Int J Comput Intell Syst, 2024, 17(1): 32

[38] Al-Ani H, Andargol L, Oh E, Chai G (2013) Categorising geotechnical properties of subsoil in surfers paradise using Geographic Information System (GIS). In Third International Conference on Geotechnique, Nagoya

[39] Mackinney W (2010) Data Structures for Statistical Computing in Python. In AQR Capital Management, LLC, Proceedings of the 9th Python in Science Conference

[40] Hastie T, Tibshirani R, Friedman J (2009) The Elements of Statistical Learing, Springer series in statistics New-York

[41] Poudel KR, Shrestha S. Urban spatial growth and geo-environment risk in Pokhara Metropolitan City: a geospatial analyis. Nepal J Environ Sci, 2025, 13115-28

[42] Pokhrel RM, Kiyota T, Kuwano R, Chiaro G, Katagiri T, Arai I. Preliminary field assessment of sinkhole damage in Pokhara, Nepal. Int J Geoengin Case Hist, 2015, 3(2): 113-125

[43] Bhattarai M, Baral S, Adhikari B. Evaluating the effectiveness of sinkhole mitigation and its impact on the local community: a case study of the Armala Sinkhole, Pokhara, Nepal. Himalayan J Appl Sci Eng, 2025

[44] Rimal B, Baral H, Stork NE, Paudyal K, Rijal S. Growing city and rapid land use transition: assessing multiple hazards and risks in the Pokhara Valley, Nepal. Land, 2015

[45] Gazette PMC (2024) 24 2 2024. [Online]. Available: https://pokharamun.gov.np/.

[46] Atamturktur S, Juang CH. Bootstrapping for characterizing the effect of uncertainty in sample statistics for braced excavations. J Geotech Geoenviron Eng, 2013, 139: 13-23

[47] Li D-Q, Tang X-S, Zhou C-B, Phoon K-K. Characterization of uncertainty in probabilistic model using bootstrap method and its application to reliability of piles. Appl Math Model, 2015, 39(17): 5310-5326

[48] Sriprasert S, Srisuradetchai P. Multi-K KNN regression with bootstrap aggregation: accurate predictions and alternative prediction intervals. Edelweiss Appl Sci Technol, 2025, 5: 2750-2764

RIGHTS & PERMISSIONS

The Author(s)