Regression-based Model Research Articles

Clinical characteristics and prognostic analysis of pediatric hemophagocytic lymphohistiocytosis using lasso-logistic regression.

This study aims to evaluate and predict mortality risks among pediatric patients with hemophagocytic lymphohistiocytosis (HLH). We conducted a retrospective analysis of pediatric patients with HLH diagnosed at the Affiliated Hospital of Zunyi Medical University between January 2012 and April 2023. Patients were divided into a death group and a survival group based on their outcomes. Risk factors for mortality were analyzed using a lasso-logistic regression model. This study included 142 pediatric patients with HLH, with a median age of 40.5 (14.75-84) months, of whom 78 (54.93%)were male. The overall mortality rate was 34.51%. Through lasso-logistic regression analysis, five independent prognostic factors were identified: concurrent central nervous system involvement, multiple organ dysfunction syndrome involving three or more organs, platelet count ≤ 42.5 × 109/L, activated partial thromboplastin time ≥ 54.05s, and the utilization of blood purification in conjunction with the HLH-94/2004 treatment protocol. The predictive value of the lasso-logistic regression model is better than that of the traditional logistic regression model (AUC: 0.906 vs 0.811, P = 0.001). Subsequently, a lasso-logistic regression-based predictive model incorporating these identified risk factors was developed. Our lasso-logistic regression-based prediction model may help to identify high-risk patients with HLH early, thereby enabling the timely initiation of appropriate treatment interventions.

Retraction Note: Automatic identification of drug sensitivity of cancer cell with novel regression-based ensemble convolution neural network model

Retraction Note: Automatic identification of drug sensitivity of cancer cell with novel regression-based ensemble convolution neural network model

Machine learning based prediction for bulk porosity and static elastic modulus of Yungang Grottoes sandstone

In this work, four mainstream machine learning (ML) techniques are used to evaluate the bulk porosity and static elastic modulus of weathered Yungang Grottoes sandstone. Datasets are gathered from the experiments, which includes 432 groups effective experimental data including 8 inputs features. bulk porosity and static elastic modulus were considered as outputs to determine the weathering degrees of Yungang Grottoes sandstone. The 4 performance criteria were used to evaluate the ML models. Results demonstrate that the Artificial Neural Network (ANN) is the best-fitted models for estimating the bulk porosity and static elastic modulus compared to Multiple Linear Regression (MLR), Support Vector Regression (SVR), Gaussian Process Regression (GPR). The accuracy of the trained model for static elastic modulus is slightly higher than that of bulk porosity. The GPR and ANN model can accurately predict the bulk porosity and static elastic modulus in training stages. The ANN with multi-hidden layers developed is competent with high degree of precision and generalization ability for bulk porosity and static elastic modulus compared to other selected regression-based ML models (MLR, SVR, and GPR). The coefficient of determinations of ANN in the range of (0.9537–0.9641) during the testing stages is more stable and higher than that of (0.8883–0.9453) other built ML models. The prediction efficiency of pretrained ANN model was well adjusted for the actual and forecast datasets at the training and testing stages, and the error range was no more than 0.7% and 0.15 GPa at both stages of prediction for bulk porosity and static elastic modulus respectively. And the ANN based static elastic modulus prediction model’s error proportions significantly decreased and were confined to a modest range between + 10% and − 10%. The proposed surrogate models are valid for the bulk porosity ranging from 7 to 14% and the static elastic modulus ranging from 0.7 to 1.4 Gpa, which can be utilized for the accurate and fast prediction of the weathering degrees of Yungang Grottoes sandstone.

A Machine Learning Assisted Non-Enzymatic Electrochemical Biosensor to Detect Urea Based on Multi-Walled Carbon Nanotube Functionalized with Copper Oxide Micro-Flowers.

Detecting urea is crucial for diagnosing related health conditions and ensuring timely medical intervention. The addition of machine learning (ML) technologies has completely changed the field of biochemical sensing, providing enhanced accuracy and reliability. In the present work, an ML-assisted screen-printed, flexible, electrochemical, non-enzymatic biosensor was proposed to quantify urea concentrations. For the detection of urea, the biosensor was modified with a multi-walled carbon nanotube-zinc oxide (MWCNT-ZnO) nanocomposite functionalized with copper oxide (CuO) micro-flowers (MFs). Further, the CuO-MFs were synthesized using a standard sol-gel approach, and the obtained particles were subjected to various characterization techniques, including X-ray diffraction (XRD), field emission scanning electron microscopy (FESEM), and Fourier transform infrared (FTIR) spectroscopy. The sensor's performance for urea detection was evaluated by assessing the dependence of peak currents on analyte concentration using cyclic voltammetry (CV) at different scan rates of 50, 75, and 100 mV/s. The designed non-enzymatic biosensor showed an acceptable linear range of operation of 0.5-8 mM, and the limit of detection (LoD) observed was 78.479 nM, which is well aligned with the urea concentration found in human blood and exhibits a good sensitivity of 117.98 mA mM-1 cm-2. Additionally, different regression-based ML models were applied to determine CV parameters to predict urea concentrations experimentally. ML significantly improves the accuracy and reliability of screen-printed biosensors, enabling accurate predictions of urea levels. Finally, the combination of ML and biosensor design emphasizes not only the high sensitivity and accuracy of the sensor but also its potential for complex non-enzymatic urea detection applications. Future advancements in accurate biochemical sensing technologies are made possible by this strong and dependable methodology.

Comparing AI/ML approaches and classical regression for predictive modeling using large population health databases: Applications to COVID-19 case prediction

BackgroundResearch comparing artificial intelligence and machine learning (AI/ML) methods with classical statistical methods applied to large population health databases is limited. ObjectivesThis retrospective cohort study aimed to compare the predictive performance of AI/ML algorithms against conventional multivariate logistic regression models using linked health administrative data. MethodsUsing Ontario's population health databases, we created a cohort of residents of the city of Ottawa, Ontario, who underwent a PCR test for COVID-19 between March 10, 2020, and May 13, 2021. Using demographic, socio-economic and health data (including COVID-19 PCR test results and available, symptom data), we developed predictive models for the purpose of COVID-19 case identification using the following approaches: classical multivariate logistic regression (LR); deep neural network (DNN); random forest (RF); and gradient boosting trees (GBT). Model performance comparisons were made using the area under the curve (AUC) swarm plot for 10-fold cross-validation. ResultsThe cohort consisted of n = 351,248 Ottawa residents tested for COVID-19 during the study period. Among whom, a total of n = 883,879 unique COVID-19 tests were performed (2.6 % positive test results). Inclusion of COVID-19 symptoms data in the analysis improved model performance and variable predictive value across all tested models (p < 0.0001), with the 10-fold cross-validation AUC increasing to near or over 0.7 in all models when symptoms data were included. In various pairwise comparisons, the GBT method had the highest predictive ability (AUC = 0.796 ± 0.017), significantly outperforming multivariate logistic regression and the other AI/ML approaches. ConclusionsConventional multivariate regression-based models are better than some and worse than other machine learning algorithms to provide good predictive accuracy in a moderate dataset with a reasonable number of features. However, whenever possible, the AI/ML GBT approach should be considered.

Flexural properties and optimisation of hybrid composites reinforced by carbon, glass and flax fibres

This paper presents a study on the flexural properties of hybrid composites (HC) reinforced by carbon, glass, and flax fibres, utilising Finite Element Analysis (FEA). Possible optimal layups are identified through the application of design rules, and regression-based models are developed to predict flexural strength. With the developed models, multi-objective optimisation is done to minimise the cost and weight subjected to a given minimum required flexural strength. By integrating numerical simulation, regression analysis, and optimisation techniques, this research offers a practical approach to the design and fabrication of HC with enhanced mechanical performance and cost-efficiency.

Risk Factors and Predictive Models for Postpartum Hemorrhage after Vaginal Delivery

Background: Postpartum hemorrhage (PPH) remains one of the biggest reasons of maternal morbidity and mortality. Clinical prediction of PPH remains challenging, particularly in the case of a vaginal birth. The purpose of this research is identifying patients at risk for PPH in vaginal delivery by using risk factors and predictive models. Methods: 1840 cases who underwent vaginal deliveries at Beijing Ditan Hospital, Capital Medical University between December 2020 to December 2022, which were divided into two groups based on the amount of blood loss (PPH and non-PPH groups). Fourteen risk factors could cause increased risk of PPH, including demographic characteristics and placental anomalies factors. Logistic regression analysis was used to influence the risk factors of PPH in vaginal delivery. According to the results of multivariate logistic regression analysis, a risk prediction model was established, the Hosmer-Lemeshow test was used to assess the model fit. Results: A total of 94 cases presented with PPH in this study, and the incidence of PPH was 5.10% (94/1840). Two items including macrosomia (odds ratio (OR): 2.229, 95% confidence interval (95% CI): 1.062–4.679) and placental anomalies (OR: 4.095, 95% CI: 2.488–6.742) were independent risk factors affecting the occurrence of PPH with vaginal delivery (p &lt; 0.05). Conclusion: The construction of a logistic regression-based model can be used to predict the risk of PPH after vaginal delivery, predictability to be studied further. Clinically, more attention should be paid to vaginal delivery, early identification and screening of high-risk factors for PPH, as well as timely preventive interventions for high-risk groups so as to reduce the risk of PPH.

Modeling of magnetocaloric effect in RE2X2Y ternary compounds for cooling technology using hybrid intelligent learning computational methods

Cooling technology that employs magnetocaloric effect of magnetic refrigerants has been lately recognized as a green system with immense potential to meet the global refrigeration demand. Hugeness of magnetocaloric impact is one of the major factors which determines the choice of magnetic refrigerants for cooling applications. This research endeavour explores and models magnetocaloric effect (MCE) in ternary system of compounds which combine rare-earth metals (RE = Tm, Er, Ho, Dy, Tb and Gd) due to the localized f-orbital electrons, transition metals (X = Co, Cu and Ni) due to their ferromagnetism near room temperature and other metals (Y = Al, Ga, In, Cd and Sn) using genetic algorithm hybridized support vector regression (GSVR) computational approach. The results of Gaussian transformation function-based hybrid model (G-MCE-GSVR) and polynomial transformation function-based hybrid model (P-MCE-GSVR) are compared with stepwise regression-based model (SR-MCE) using different performance metrics such as mean absolute error (MAE), correlation coefficient (CC) and root mean square error (RMSE). G-MCE-GSVR model outperforms P-MCE-GSVR model using testing data samples with improvement of 66.93 % for RMSE metric, 5.69 % for CC metric and 68.58 % for MAE metric. Similarly, the P-MCE-GSVR model performs better than SR-MCE with improvements of 60.36 %, 37.63 % and 63.41 % corresponding to performance metrics RMSE, CC and MAE, respectively. G-MCE-GSVR model was also employed for investigating the influence of applied magnetic field on magnetocaloric effect in system of RE2X2Y ternary materials. The outstanding performance demonstrated by the developed models coupled with the obtained huge magnetocaloric effect would be of great significance in strengthening practical implementation of magnetic cooling technology for addressing energy crisis and global refrigeration demand.

Regression-based model predicting cost contingencies for road network projects

Cost contingencies represent a significant markup percentage for construction projects. Many construction projects experience cost overruns due to unexpected events before or during the construction phase. Determining the cost contingency for construction projects is one of the most frequent and widespread financial problems many countries face. Since the amount included in the estimates is always insufficient to absorb the actual cost overruns of the project. The main objective of this paper is to develop a simple realistic regression model for predicting the appropriate contingency amount for road network projects before the bidding stage. The regression model has been developed based on data from real case studies for 75 road network projects to adequately consider all potential risks may face the project. Good-of-fit tests were used to select the best-fitting statistical distribution for a given dataset. Based on the analysis of selected data, the log-logistic probability distribution function is the best-fit curve representing cost overrun behavior, and the cumulative distribution function has been used to determine the realistic probability of incurred cost overrun. Based on the analysis of real case studies, it is suggested that an average of 59.722% be used as a cost overrun contingency amount for road network projects. The study outcomes will help practitioners and researchers improve future accuracy/calculations of contingency amount for road construction projects and improve the decision-making process.

Assessment of ridge regression-based machine learning model for the prediction of automotive sales based on the customer requirements

Assessment of ridge regression-based machine learning model for the prediction of automotive sales based on the customer requirements

Comparison of three artificial intelligence methods for predicting 90% quantile interval of future insulin sensitivity of intensive care patients

Insulin dosing of hyperglycemic patients in the intensive care unit (ICU) is a complex and nonlinear clinical control problem. Recent model-based glycemic control protocols predict a patient-specific and time-specific future insulin sensitivity distribution, which defines the future patient state in response to insulin and nutrition inputs. The prediction methods provide a 90% confidence interval for a future insulin sensitivity distribution for a given time horizon, making the prediction problem more specific compared to common prediction problems where the aim is to predict the expected value of the given stochastic parameter. This study proposes three alternative artificial intelligence-based insulin sensitivity prediction methods to improve the prediction accuracy and make prediction parameters better fit the clinical requirements. The proposed prediction methods use different neural network models: a classification deep neural network model, a Mixture Density Network model, and a Quantile Regression-based model. A large patient data set was used to create the neural network models, including 2357 patients and 92646 blood glucose measurements from three clinical sites (Christchurch, New Zealand, Gyula, Hungary, and Liege, Belgium). Prediction accuracy was assessed by statistical metrics expressing clinical requirements, as well as via validated in-silico virtual patient simulations comparing the clinical performance of a proven glycaemic control protocol using the alternative prediction methods to assess impact on glycemic control performance and thus the need for these alternative models.

Inferred attractiveness gravity-based models for estimating realized access at rural hospitals

Operating obstetric units in rural America is financially challenging in part due to low birth volume. Birth volume at a hospital decreases when birthers bypass it to go to a farther hospital. Beyond financial considerations, it is important from a healthcare equity perspective for hospitals to know whether certain subgroups of birthers avoid utilizing the hospital’s services. This can better inform resource allocation decisions targeting those subgroups. In this paper, we use a nonlinear programming optimization model, inferred attractiveness gravity-based model (GBM), to estimate realized access to obstetric care at hospitals in Montana. We compare three variations of GBM and benchmark our results to a regression-based conditional logit model. Results indicate that hospital attractiveness varies across the level of obstetric care provided and depends on the subgroup of birthers considered. While all GBMs produced smaller errors for hospitals with higher birth volumes, our novel variant was more accurate for low-volume hospitals. Bootstrapping analyses and resolving the models for population subgroups indicated large variations in hospital attractiveness. Research findings contribute to new knowledge about equity in access to obstetric care, the importance of considering population heterogeneity in GBMs, and the benefit of using hospital demand-based thresholds for GBMs in rural settings.

Machine learning regression-based prediction model for the autonomous control of coagulant dosing in smart water purification plants

ABSTRACT The global issue of water scarcity is escalating due to urbanization and increased demand. This paper proposes a machine learning (ML) regression-based model for automatic coagulant dosing control in smart water purification plants (SWPPs).1 The model uses random forest (RF), light gradient boosting machine (LGBM), extreme gradient boosting (XGB), and k-nearest neighbors (KNN) algorithms. Performance metrics include MAE, MSE, RMSE, MAPE, and R2. The RF algorithm showed superior performance, with MAE of 0.005, MSE of 0.002, RMSE of 0.05, and MAPE of 0.000 for anion-poly aluminum chloride dosing, and MAE of 0.007, MSE of 0.00, RMSE of 0.02, and MAPE of 0.000 for Polymax dosing. The RF model's performance is due to its robust handling of large datasets and ensemble learning approach. Limitations include testing only two coagulants and reliance on historical data. These findings suggest integrating advanced water management, energy systems, and facility management to make SWPPs feasible and efficient, establishing a foundation for future applications of ML in chemical processes.

Analysis of high temperature and strain amplitude effects on low cycle fatigue behavior of pitting corroded killed E350 BR structural steel

High-tension structural steels are prone to accelerated fatigue damage from pitting corrosion and high-temperature. Despite adverse effects, research on their low cycle fatigue (LCF) behavior is limited. Specifically, studies analyzing temperature-dependent pit sensitivity effects, considering pit-related material’s susceptibility to surface topographic features variation and stress concentration are lacking. This study conducts LCF tests on pitting corroded killed E350 BR structural steel at multiple strain amplitudes and high temperatures. It develops temperature-dependent parameters, such as the cyclic softening pit sensitivity factor, suitable for integration into existing approaches like total cyclic plastic strain energy density (CPSED), power-law, average strain energy density (SED), Coffin-Manson, and pit stress intensity factor (pit-SIF). Further, it proposes multiple linear regression-based prediction models relating total CPSED and average SED with strain amplitude and temperature. Corroded specimens show higher plastic deformation and reduced peak stress, fatigue life, and total CPSED compared to uncorroded ones. The developed parameters, integrated with average SED approach, predicts LCF life within an error band of ±1.5, while power-law relationship reduces it to ±1.2. Moreover, pit-SIF approach estimates fatigue life within an error band of ±1.5. The findings provide critical knowledge for enhanced component design, leading to structural safety, performance, and fire resilience.

Multi-environment evaluation of Clary sage (Salvia sclarea L.) selections for yield and essential oil traits under Western Himalayan conditions

Clary sage is an important aromatic herb that has a high potential for commercial cultivation under the diversified climates of the Western Himalayas. The present study was carried out to understand the performance of genotypes and their interaction with the varied environments prevalent in the Western Himalayas and to identify stable and best-performing genotypes. In the present experiment, ten superior selections were evaluated in four different environments as multi-location trials in a randomized block design with three replications during 2021 and 2022. The pooled analysis of variance revealed significant variations for genotypes, environments, and genotype × environment interaction for all the studied traits. The highest mean performance suggests that Env-1 has the most suitable conditions for clary sage cultivation. The genotypes CSIR-IHBT-SS-07 and CSIR-IHBT-SS-09 were found to be the best performers for fresh inflorescence weight (418.73 g/plant) and essential oil content (428.91 mg/Kg), respectively. The results of Eberhart and Russell’s regression-based model confirm that the genotype CSIR-IHBT-SS-07 is a stable and superior genotype for the economically important traits. Further, the gas-chromatography mass spectroscopic characterization of essential oil disclosed that CSIR-IHBT-SS-07 was unique in terms of the highest sclareol content (23.39–45.17 %) across all the test environments. Genotype + Genotype × Environment biplot analysis confirmed that all the test environments form a single mega-environment. The study is a pioneer in unraveling the stability response of clary sage in Western Himalaya and provides a strong base for the identification of superior genotypes and selection strategy for future genetic improvement programs of clary sage.

Real-time solar PV generation in a building using LSTM-based time series forecasting

The increasing demand for renewable energy has forced the researchers to lay importance on studying the behavior of these sources by monitoring the operational data and performing data analysis. The application of data analytics to renewable energy is one of the most powerful methods to maintain the system's reliability and stability. As the power system becomes more complex the necessity to perform data analytics also increases. When data analytic techniques are applied to solar energy generations through Photovoltaic (PV) dataset, the possible behavior of PV generation performance which is affected by changes in environmental conditions can be predicted and further analytical approaches allow us to detect possible PV panel and inverter failures. This paper is an attempt towards applying the intelligent data analytics approaches to solar PV generation of a real-time photovoltaic plant. The main purpose of the data analytics platform is to analyze the energy yield, assess the PV system performance, and forecast solar PV generation. In this article, Long Short-Term Memory (LSTM) machine learning model is developed to assess and interpret the available information from the gathered data of the PV plant. The proposed model is a sequence-to-sequence regression-based forecasting model which performs time series forecasting in a real-time 100 kW PV plant in the University building of SRM Institute of Science and Technology, Kattankulathur, Tamilnadu.

Fusion Logistic Regression: Advancing Towards Stroke Early Warning

In this study, we focus on developing a logistic regression-based binary classification model to identify potential hemorrhagic stroke patients accurately. The model utilizes personal history, medical records, onset and treatment information from a training set of 100 hemorrhagic stroke patients. The objective is to predict the extent of hematoma expansion, assessing whether patients face significant health risks. Compared to traditional methods for identifying hemorrhagic strokes, such as decision trees, support vector machines, random forests, and gradient boosting machines, our logistic regression model demonstrates significant advantages in performance metrics such as F1 score and recall. Moreover, it achieves an accuracy rate of 96% in testing, surpassing other comparative models. Additionally, the model provides each patient with precise disease probability predictions, aiding in early treatment, alleviating economic burdens, and substantially reducing the technical complexity for medical professionals during the diagnosis and treatment process.

Does change in area-level deprivation, change health outcomes? A latent class growth analysis of population data.

BackgroundWhile the link between deprivation and health is established, little is known about the association of persistent deprivation or deprivation trajectories over time with adverse health outcomes. AimTo examine latent class trajectories of area level-deprivation from 2010-2016 and their association with social distress indicators from 2017-2021. MethodAnalysis was based on linked administrative data (N=1,569,110) from NI GP Registrations Index, emergency department records (2017-21), enhanced prescribing data (2017-21) and biennial area-level deprivation (2010-2016). Latent class growth analysis (LCGA) identified deprivation trajectories between 2010 and 2016; and regression-based models, adjusted for age, sex and urbanicity, examined their association with indicators of individual level distress: occurrence of death, any accident and emergency (A&amp;E) visit, and any psychotropic medication between 2017 and 2021. ResultsBased on comparison of models and fit statistics, a non-linear LGCA model identified six trajectories: consistently deprived (21.2%); slight improvement (7.1%); substantial improvement (1.2%); consistently average deprivation (31.7%); slight worsening (18.5%); and consistently non-deprived (20.4%). Consistently living in a deprived area was associated with higher likelihood of death, A&amp;E visits and receipt of hypnotics, anxiolytics, and anti-depressants. Compared to consistent deprivation, slight improvement was associated with lower likelihood of death (OR=0.94, CI: 0.91-0.97) and all psychotropic medication in the follow-up period. ImplicationsFindings indicate that even a slight improvement in deprivation status can have a positive influence on health outcomes. Further population-wide analysis of deprivation trajectories, their determinants and associated health impact will inform targeted interventions that promote social mobility and improved health.

Derivation of Indices of Cognitive Change Among Hispanic Adults and Elders.

Determining the influence of race and ethnicity on change in cognitive test performance has significant implications for clinical practice and research in populations at risk for Alzheimer disease. To evaluate the significance of race and ethnicity in predicting longitudinal cognitive test performance and to develop models to support evidence-based practice. This prognostic study included baseline and 24-month follow-up data that were obtained from the Health and Aging Brain Study-Health Disparities (HABS-HD) study, an ongoing longitudinal observational study of aging and dementia in a multiracial, multiethnic cohort. Participants included community-dwelling adults and elders living in the Dallas and Fort Worth metropolitan area who were Hispanic and non-Hispanic adults older than the age of 50 years and were cognitively unimpaired. The primary exposure of interest was time, measured in months. Demographic variables included age, sex, education, and race and ethnicity. Cognitive domains included attention and working memory, processing speed, language, memory, and executive functioning. Linear regression models predicted follow-up performance from baseline performance and demographic variables for 13 commonly used neuropsychological tests. Follow-up testing was the primary outcome for all domains. Raw scores from 13 standardized tests were used for analyses. This study included 799 adults who were cognitively unimpaired (352 Hispanic individuals [44.1%]; 447 non-Hispanic individuals [55.9%]; 524 female [65.6%]; mean [SD] age, 65.4 [8.1] years). In the regression models, all 13 follow-up scores were significantly predicted from their respective baseline scores and demographic variables. Baseline performance and education were the most consistent predictors of follow-up scores, contributing to all 13 models. Age was significantly associated with follow-up in 11 models, and sex was significant in 5 models. Race and ethnicity contributed to 10 of 13 models, with Hispanic participants predicted to have poorer follow-up scores than their non-Hispanic White counterparts on each test. In this longitudinal study of cognitive change in Hispanic and non-Hispanic older adults who were cognitively unimpaired, standardized regression-based models were influenced by multiple demographic variables, including race and ethnicity. These findings highlight the importance of including race and ethnicity in such cognitive change models. This ability to accurately predict cognitive change is expected to become increasingly important as clinical practice and clinical trials need to become more diverse and culturally appropriate in this burgeoning global medical and societal crisis.

Development of machine learning-based standalone GUI application for predicting hydraulic conductivity and compaction parameters of lateritic soils

Hydraulic conductivity and compaction parameters are the key considerations in selecting lateritic soils for engineering construction. Nevertheless, the complexity and high cost of the required tests have driven many contractors and field engineers to skip them, resulting in a succession of engineering structure failures. To overcome this limitation, this study developed machine learning-based standalone GUI application to predict lateritic soils’ hydraulic conductivity (K), maximum dry density (MDD) and optimum moisture content (OMC) from indices including specific gravity, liquid limit, plasticity index, linear shrinkage and fine content. To achieve this goal, the geotechnical properties of three hundred samples, collected using grid sampling method from thirty different lateritic deposits in southwestern Nigeria, were evaluated through laboratory tests. The test results were used to train predictive models using artificial neural network (ANN), adaptive neuro-fuzzy inference system (ANFIS) and Gaussian process regression (GPR). The models’ performance was compared using coefficient of determination (R2), root mean squared error (RMSE), mean absolute percentage error (MAPE) and mean absolute error (MAE). Based on these performance metrics, ANN demonstrated the best performance (R2 = 0.9835, 0.9797, 0.9999; RMSE = 7.938, 0.252, 2.09E-08; MAPE = 0.288, 1.114, 1.587; MAE = 5.432, 0.169, 1.1E-08) for MDD, OMC and K, respectively, followed by GPR and then ANFIS. Thus, the ANN models were selected and embedded in a standalone GUI application to enhance easy and quick prediction of lateritic soils’ MDD, OMC and K. The validity of the ANN-based standalone GUI application was demonstrated by comparing it favorably to notable regression-based models in the literature.