Artificial Neural Network Regression Research Articles

An Investigation of Slurry Erosion Behaviour in Plasma-Sprayed Carbon Nanotube-Reinforced Fly Ash/Alumina Coatings Using Experimental Analysis and Artificial Neural Computing for Marine and Offshore Applications

This study investigates carbon nanotube (CNT)-reinforced alumina fly ash (FA) coatings, namely AF (unreinforced), 1CAF (with 1wt% CNT), and 2CAF (with 2wt% CNT), on marine-grade steel. Microstructural analysis shows 1CAF coatings denser by ~15.32% due to CNT reinforcement, while 2CAF coatings display ~9.68% increased porosity from CNT agglomeration. Raman spectroscopy confirms CNT retention. 1CAF coatings exhibit ~14.66% higher microhardness, ~15.96% higher adhesion strength, and ~15.66% improved fracture toughness compared to AF coatings, attributed to pore sealing through CNT reinforcement. Enhanced erosion resistance (~14.59%) in 1CAF coatings was observed due to improved mechanical properties and CNTs mitigating crack propagation. Validation through an artificial neural network (ANN) modeling and regression analysis supports 1CAF coatings’ promise for harsh marine environments, offering enhanced durability.

Soft Computing Modeling Including Artificial Neural Network, Non-linear, and Linear Regression Models to Predict the Compressive Strength of Sustainable Mortar Modified with Palm Oil Fuel Ash

Soft Computing Modeling Including Artificial Neural Network, Non-linear, and Linear Regression Models to Predict the Compressive Strength of Sustainable Mortar Modified with Palm Oil Fuel Ash

Research on Decoupling Model of Six-Component Force Sensor Based on Artificial Neural Network and Polynomial Regression.

A two-stage decoupling model based on an artificial neural network with polynomial regression is proposed for the six-component force sensor load decoupling problem in the case of multidimensional mixed loading. The six-dimensional load categorization stage model constructed in the first stage combines 63 load category label sets with a deep BP neural network. The six-dimensional load regression stage model was constructed by combining polynomial regression with a BP neural network in the second stage. Meanwhile, the six-component force sensor with a fiber Bragg grating (FBG) sensor as the sensitive element was designed, and the elastomer simulation and calibration experimental dataset was established to realize the validation of the two-stage decoupling model. The results based on the simulation data show that the accuracy of the classification stage is 93.65%. The MAPE for the force channel in the regression stage is 6.29%, and 3.24% for the moment channel. The results based on experimental data show that the accuracy of the classification stage is 87.80%. The MAPE for the force channel in the regression phase is 5.63%, and 4.82% for the moment channel.

Machine learning-based Nusselt number prediction for falling-film evaporators in absorption refrigeration systems

The evaporator serves as a pivotal component of an absorption refrigeration system (ARS), bolstered by improved wettability and mitigation of the irregular arrangement of the liquid film. The complexities of the evaporation heat transfer characteristics of falling films have led to numerous empirical correlations for falling-film Nusselt number (Nu) estimation, demanding substantial computational resources and experimental expenses and limiting their universal applicability. This study used machine learning to predict Nu of a falling-film evaporator. Artificial neural network (ANN) and random forest regression (RFR) models were used to predict heat transfer characteristics and establish a novel Nu correlation using an experimentally derived dataset. Correlations of falling-film Nu were established for six tube types using empirical data, ANN, and RFR predictions and then compared based on experimental falling-film Nu. The ANN and RFR provided high prediction performance within 10 % error and an R2 of 0.985 and 0.999, respectively. The Nu correlations derived from ANN and RFR demonstrated satisfactory agreement, typically within ±15 % error of experimental results lower than that of the empirical correlations. The optimized ANN and RFR demonstrates their promising ability to predict heat transfer characteristics and establish Nu correlations, thereby reducing the need for immense experimental efforts.

Hyperspectral imaging predicts free fatty acid levels, peroxide values, and linoleic acid and oleic acid concentrations in tree nut kernels

Imaging technologies are advancing rapidly in food processing systems to reduce food waste. However, modelling techniques, sample sizes and dataset proportioning methods significantly affect the performance of models in predicting food quality variables. This study examined the potential of hyperspectral imaging to predict peroxide values (PV), free fatty acid (FFA) levels and fatty acid concentrations, including oleic and linoleic acid, in two tree nuts, canarium and macadamia. The effectiveness of artificial neural network (ANN) regression and partial least squares regression (PLSR) techniques to make these predictions was examined. Additionally, the importance of the dataset size on prediction accuracy and the dataset-proportioning method for developing predictive models were assessed. Both ANN and PLSR models predicted FFA levels and oleic acid concentrations with high accuracy, but PV and linoleic acid concentrations were predicted poorly. Changing the test-dataset proportioning method for the small dataset led to comparable R2test values by both ANN and PLSR in predicting FFA levels. Successful prediction of FFA levels could be explained partly by high variability, even in the dataset with the small number of samples. The study highlights the significance of factors like dataset size, test-dataset proportioning method, and model selection when predicting the quality attributes.

Machine learning models for predicting thermal desorption remediation of soils contaminated with polycyclic aromatic hydrocarbons

Among various remediation methods for organic-contaminated soil, thermal desorption stands out due to its broad treatment range and high efficiency. Nonetheless, analyzing the contribution of factors in complex soil remediation systems and deducing the results under multiple conditions are challenging, given the complexities arising from diverse soil properties, heating conditions, and contaminant types. Machine learning (ML) methods serve as a powerful analytical tool that can extract meaningful insights from datasets and reveal hidden relationships. Due to insufficient research on soil thermal desorption for remediation of organic sites using ML methods, this study took organic pollutants represented by polycyclic aromatic hydrocarbons (PAHs) as the research object and sorted out a comprehensive data set containing >700 data points on the thermal desorption of soil contaminated with PAHs from published literature. Several ML models, including artificial neural network (ANN), random forest (RF), and support vector regression (SVR), were applied. Model optimization and regression fitting centered on soil remediation efficiency, with feature importance analysis conducted on soil and contaminant properties and heating conditions. This approach enabled the quantitative evaluation and prediction of thermal desorption remediation effects on soil contaminated with PAHs. Results indicated that ML models, particularly the RF model (R2 = 0.90), exhibited high accuracy in predicting remediation efficiency. The hierarchical significance of the features within the RF model is elucidated as follows: heating conditions account for 52 %, contaminant properties for 28 %, and soil properties for 20 % of the model's predictive power. A comprehensive analysis suggests that practical applications should emphasize heating conditions for efficient soil remediation. This research provides a crucial reference for optimizing and implementing thermal desorption in the quest for more efficient and reliable soil remediation strategies.

Development of a magnetic nanocomposite sorbent (NiCoMn/Fe3O4@C) for efficient extraction of methylene blue and Auramine O

A rapid and efficient method for the simultaneous monitoring and recovery of Auramine O (AO) and Methylene Blue (MB) dyes from water samples is presented. This method, named ultrasound-assisted dispersive-magnetic nanocomposites-solid-phase microextraction (UA-DMN-μSPE), utilizes NiCoMn/Fe3O4@C composite sorbents. Response surface methodology (RSM) combined with artificial neural networks (ANN) and generalized regression artificial neural network (GRNN) under central composite design (CCD) was employed to optimize various parameters for efficient extraction, followed by further refinement using desirability function analysis (DFA) and genetic algorithms (GA). Under optimized conditions, the method achieved exceptional recovery rates (99.5 ± 1.2% for AO and 99.8 ± 1.1% for MB) with acetone as the eluent. Additionally, a high preconcentration factor of 45.50 and 47.30 for AO and MB, respectively, was obtained. Low detection limits of 0.45 ng mL⁻1 (AO) and 1.80 ng mL⁻1 (MB) were achieved with wide linear response ranges (5–1000 and 5–2000 ng mL⁻1 for AO and MB, respectively). The method exhibited good stability with RSDs below 3% for five recycling runs, and minimal interference from various ions was observed. This UA-DMN-μSPE-UV/Vis method offers significant advantages in terms of efficiency, preconcentration, and detection limits, making it a valuable tool for the analysis of AO and MB in water samples.

Predicted risk factors associated with secondary infertility in women: univariate and multivariate logistic regression analyses.

Infertile women are those who regularly engage in unprotected intercourse for a period of at least 1 year and are unable to become clinically pregnant. Primary infertility means the inability of couples to conceive, without any previous successful pregnancies. Secondary Infertility refers to the inability to get pregnant for 12 months, after having a previous pregnancy for one time at least. The objectives of the current study were to analyze risk factors for secondary infertility and compare the predictive accuracy of artificial neural network (ANN) and multiple logistic regression models. The study was conducted at The University Institute of Public Health collecting data from Gilani Ultrasound Center 18 months after approval of synopsis. A total of 690 women (345 cases and 345 controls) were selected. The women selected for the case group had to be 20-45 years of age, had any parity, and had a confirmed diagnosis of secondary infertility. Multiple logistic regression (MLR) and ANN were used. The chance of secondary infertility was 2.91 times higher in women living in a joint family [odds ratio (OR) = 2.91; 95% confidence interval (CI) (1.91, 4.44)] and was also 2.35 times higher for those women who had relationship difficulties with their husband [OR = 2.35; 95% CI (1.18, 4.70)]. Marriage at an earlier age was associated with secondary infertility with β being negative and OR being < 1 [OR = 0.94; 95% CI (0.88, 0.99)]. For the logistic regression model, the area under the receiver operative characteristic curve (ROC) was 0.852 and the artificial neural network was 0.87, which was better than logistic regression. Identified risk factors of secondary infertility are mostly modifiable and can be prevented by managing these risk factors.

A study on nitrogen oxide emission prediction in Taichung thermal power plant using artificial intelligence (AI) model

In recent studies, artificial intelligence models have been developed for the prediction of nitrogen oxide emissions from thermal power plants. This research utilizes an artificial intelligence prediction model with more coal input features than boiler features, including volatility, ash content, sulfur content, fixed carbon, total moisture, calorific value, grinding rate, fuel ratio, coal feeding rate, boiler efficiency, total air volume, and excess air volume. The paper delves into the importance analysis of input features for artificial intelligence models. Moreover, feature importance analysis is not only a prerequisite for predicting nitrogen oxide emissions but also a study providing insights into model performance. An artificial neural networks (ANN) regression model is employed to predict nitrogen oxide emissions, and the results demonstrate that the number of feature importance significantly impacts model performance. The best model performance is achieved with eight specific input features. Finally, after the training and validation processes, the ANN model yields the optimal coefficient of determination (R2) value.

Real‐time monitoring and ageing detection algorithm design with application on SiC‐based automotive power drive system

AbstractThe article describes an innovative methodology for the design and experimental validation of monitoring and anomaly detection algorithms, with a particular focus on the aging phenomenon, linked to the anomalous modification of the , in devices switching in power electronic systems integrated into modern high‐performance electrified vehicles. The case study concerns an electric drive for fully electrified vehicles, in which a three‐phase axial flux synchronous motor integrated into a wheel motor (Elaphe) is used and in which a high‐efficiency three‐phase inverter, designed with SiC technology (silicon carbide). The article proposes the design and validation of the innovative aging monitoring and detection system, in four consecutive phases. The first phase involves the creation of a real‐time model of electric drive, validated through experimental data extrapolated directly during a WLTP (Worldwide Harmonized Light Vehicle Test Procedure) test. The second phase consists of the creation of a virtual dataset representative of the aging phenomenon, via an anomaly injection procedure, emulating this phenomenon with a scaling factor (depending on the value of the ) on the current phase of the motor, relating to the inverter branch whose SiC device is affected. The third phase concerns the design of an estimator of the , based on an ANN (Artificial Neural Network) regression model, and involves a data manipulation phase with features extraction and reduction techniques. The fourth and final phase, involves the experimental validation of the method, through PIL (Processor‐In‐the‐Loop) tests, integrating the monitoring algorithm (consisting of a real‐time model and AI‐based regression model) on the NXPs32k144 embedded platform (based on Cortex‐M4), making the algorithm interact with the electric drive model on which anomaly injection is applied.

Efficient removal of ciprofloxacin and ofloxacin from aqueous solutions using a novel nano-scale adsorbent: Modeling, optimization, and characterization

In the fascinating realm of water purification, our study unveils the remarkable potential of a cutting-edge nano-scale adsorbent—combining graphene oxide (GO), chitosan (CS), and polydopamine (PDA)—in efficiently remove ciprofloxacin (CPF) and ofloxacin (OFL) from aqueous solutions. Our exploration delves deep into the adsorbent's character, utilizing a range of analytical techniques including SEM, RAMAN, FTIR, TGA, BET, XRD, and Zeta potential analyses provided insights into the adsorbent's properties. Modeling the adsorption process with Response Surface Methodology (RSM), Artificial Neural Network (ANN) and General Regression Neural Network (GRNN) indicated excellent predictions by GRNN, with RMSE = 0.0200 and 0.0166, MAE = 0.0082 and 0.0092, as well as AAD = 0.0002 and 0.0006, highlighting its modeling power.Optimization using genetic algorithm (GA) revealed maximum CPF removal efficiency of approximately 95.20% under pH = 6.3, sonication time = 9.0 min, adsorbent dosage = 2.10 g L⁻1, temperature = 45 °C and initial CPF concentration = 90.0 mg L⁻1. Similarly, OFL removal reached about 95.50% under pH = 6.30, sonication time = 8.0 min, adsorbent dosage = 2.0 g L⁻1, temperature = 45 °C and OFL concentration = 115.0 mg L⁻1. RSM optimization closely aligned with GA results. Pseudo-second-order (PSO) kinetic model and Langmuir isotherm model best fitted the experimental data for both antibiotics. Thermodynamic analysis indicated a favorable and spontaneous adsorption process for CPF and OFL. The study concludes that the proposed adsorbents show effectiveness in removing CPF and OFL at lower doses and shorter sonication times compared to various reported adsorbents.

Valence-isomer selective cycloaddition reaction of cycloheptatrienes-norcaradienes

The rapid and precise creation of complex molecules while controlling multiple selectivities is the principal objective in synthetic chemistry. Combining data science and organic synthesis to achieve this goal is an emerging trend, but few examples of successful reaction designs are reported. We develop an artificial neural network regression model using bond orbital data to predict chemical reactivities. Actual experimental verification confirms cycloheptatriene-selective [6 + 2]-cycloaddition utilizing nitroso compounds and norcaradiene-selective [4 + 2]-cycloaddition reactions employing benzynes. Additionally, a one-pot asymmetric synthesis is achieved by telescoping the enantioselective dearomatization of non-activated benzenes and cycloadditions. Computational studies provide a rational explanation for the seemingly anomalous occurrence of thermally prohibited suprafacial [6 + 2]-cycloaddition without photoirradiation.

Using statistical and machine learning approaches to describe estuarine tidal dynamics

Abstract Estuaries are ecologically valuable regions where tidal forces move large volumes of water. To understand the ongoing physical processes in such dynamic systems, a series of estuarine monitoring stations is required. Based on the measurements, estuarine dynamics can be described by key values, so-called tidal characteristics. The reconstruction and prediction of tidal characteristics by suitable approaches is essential to discover natural or anthropogenic changes. Therefore, it is of interest to inter- and extrapolate measured values in time and to investigate the spatial relationship between different stations. Normally, such system analyses are performed by deterministic numerical models. However, to facilitate long-term investigations also, statistical and machine learning approaches are good options. For a Weser estuary case study, we implemented three approaches (linear, non-linear, and artificial neural network regression) with the same database to enable the prediction of tidal extrema. Thereby we achieve an accuracy of 0.4–2.5% derivation (based on the RMSEs) while approximating measured values over 19 years. This proves that the approaches can be used for hindcast studies as well as for future analysis of system changes. Our work can be understood as a proof of concept for the practical potential of neural networks in estuarine system analysis.

A comparative study of 11 non-linear regression models highlighting autoencoder, DBN, and SVR, enhanced by SHAP importance analysis in soybean branching prediction.

To explore a robust tool for advancing digital breeding practices through an artificial intelligence-driven phenotype prediction expert system, we undertook a thorough analysis of 11 non-linear regression models. Our investigation specifically emphasized the significance of Support Vector Regression (SVR) and SHapley Additive exPlanations (SHAP) in predicting soybean branching. By using branching data (phenotype) of 1918 soybean accessions and 42k SNP (Single Nucleotide Polymorphism) polymorphic data (genotype), this study systematically compared 11 non-linear regression AI models, including four deep learning models (DBN (deep belief network) regression, ANN (artificial neural network) regression, Autoencoders regression, and MLP (multilayer perceptron) regression) and seven machine learning models (e.g., SVR (support vector regression), XGBoost (eXtreme Gradient Boosting) regression, Random Forest regression, LightGBM regression, GPs (Gaussian processes) regression, Decision Tree regression, and Polynomial regression). After being evaluated by four valuation metrics: R2 (R-squared), MAE (Mean Absolute Error), MSE (Mean Squared Error), and MAPE (Mean Absolute Percentage Error), it was found that the SVR, Polynomial Regression, DBN, and Autoencoder outperformed other models and could obtain a better prediction accuracy when they were used for phenotype prediction. In the assessment of deep learning approaches, we exemplified the SVR model, conducting analyses on feature importance and gene ontology (GO) enrichment to provide comprehensive support. After comprehensively comparing four feature importance algorithms, no notable distinction was observed in the feature importance ranking scores across the four algorithms, namely Variable Ranking, Permutation, SHAP, and Correlation Matrix, but the SHAP value could provide rich information on genes with negative contributions, and SHAP importance was chosen for feature selection. The results of this study offer valuable insights into AI-mediated plant breeding, addressing challenges faced by traditional breeding programs. The method developed has broad applicability in phenotype prediction, minor QTL (quantitative trait loci) mining, and plant smart-breeding systems, contributing significantly to the advancement of AI-based breeding practices and transitioning from experience-based to data-based breeding.

Multivariable predictive models for the estimation of power consumption (kW) of a Semi-autogenous mill applying Machine Learning algorithms [Modelos predictivos multivariables para la estimación de consumo de potencia (kW) de un molino Semi - autógeno aplicando algoritmos de Machine Learning]

This research aimed to develop machine learning (ML) models to estimate power consumption (Kw) in a Semi-autogenous mill in the mining industry. Using Machine Learning algorithms considering various operating variables for the different models such as Multiple Linear Regression (RLM), Decision Tree Regression (RAD), Random Forest Regression (RBA) and Regression Artificial Neural Networks (ANN). The methodology adopted was applied, with an experimental design with a descriptive and transversal approach. The results of the application of these models revealed significant differences in terms of predictive efficiency. The RLM and RRNA stood out with coefficients of determination (R²) of 0.922 and 0.939, respectively, indicating a substantial capacity to explain the variability in power consumption. In contrast, the tree-based models (RAD and RBA) showed inferior performance, with R² of 0.762 and 0.471. When analyzing key metrics such as Mean Absolute Error (MAE), Mean Square Error (MSE) and Root Root Mean Square Error (RMSE), it was confirmed that both RLM and RRNA outperformed the tree-based models. These results support the choice of RLM and RRNA as preferred models for estimating power consumption in a Semi-autogenous mill.

Comparative Analysis of Machine Learning Algorithms Evaluating the Single Nucleotide Polymorphisms of Metabolizing Enzymes with Clinical Outcomes Following Intravenous Paracetamol in Preterm Neonates with Patent Ductus Arteriosus.

Pharmacogenomics has been identified to play a crucial role in determining drug response. The present study aimed to identify significant genetic predictor variables influencing the therapeutic effect of paracetamol for new indications in preterm neonates. Paracetamol has recently been preferred as a first-line drug for managing Patent Ductus Arteriosus (PDA) in preterm neonates. Single Nucleotide Polymorphisms (SNPs) in CYP1A2, CYP2A6, CYP2D6, CYP2E1, and CYP3A4 have been observed to influence the therapeutic concentrations of paracetamol. The purpose of this study was to evaluate various Machine Learning Algorithms (MLAs) and bioinformatics tools for identifying the key genotype predictor of therapeutic outcomes following paracetamol administration in neonates with PDA. Preterm neonates with hemodynamically significant PDA were recruited in this prospective, observational study. The following SNPs were evaluated: CYP2E1*5B, CYP2E1*2, CYP3A4*1B, CYP3A4*2, CYP3A4*3, CYP3A5*3, CYP3A5*7, CYP3A5*11, CYP1A2*1C, CYP1A2*1K, CYP1A2*3, CYP1A2*4, CYP1A2*6, and CYP2D6*10. Amongst the MLAs, Artificial Neural Network (ANN), C5.0 algorithm, Classification and Regression Tree analysis (CART), discriminant analysis, and logistic regression were evaluated for successful closure of PDA. Generalized linear regression, ANN, CART, and linear regression were used to evaluate maximum serum acetaminophen concentrations. A two-step cluster analysis was carried out for both outcomes. Area Under the Curve (AUC) and Relative Error (RE) were used as the accuracy estimates. Stability analysis was carried out using in silico tools, and Molecular Docking Studies (MDS) were carried out for the above-mentioned enzymes. Two-step cluster analyses have revealed CYP2D6*10 and CYP1A2*1C to be the key predictors of the successful closure of PDA and the maximum serum paracetamol concentrations in neonates. The ANN was observed with the maximum accuracy (AUC = 0.53) for predicting the successful closure of PDA with CYP2D6*10 as the most important predictor. Similarly, ANN was observed with the least RE (1.08) in predicting maximum serum paracetamol concentrations, with CYP2D6*10 as the most important predictor. Further MDS confirmed the conformational changes for P34A and P34S compared to the wildtype structure of CYP2D6 protein for stability, flexibility, compactness, hydrogen bond analysis, and the binding affinity when interacting with paracetamol, respectively. The alterations in enzyme activity of the mutant CYP2D6 were computed from the molecular simulation results. We have identified CYP2D6*10 and CYP1A2*1C polymorphisms to significantly predict the therapeutic outcomes following the administration of paracetamol in preterm neonates with PDA. Prospective studies are required for confirmation of the findings in the vulnerable population.

Risk factor analysis and prediction model construction for surgical patients with venous thromboembolism: A prospective study

Abstract Objective: Patients undergoing surgery are at high risk of developing venous thromboembolism (VTE). This study aimed to determine the predictive value of risk factors for VTE in surgical patients and to develop a prediction model by integrating independent predictors. Methods: A total of 1111 patients who underwent surgery at clinical departments in a tertiary general hospital were recruited between May and July 2021. Clinical data, including patient-related, surgery-related, and laboratory parameters, were extracted from the hospital information system and electronic medical records. A VTE prediction model incorporating ten risk variables was constructed using artificial neural networks (ANNs). Results: Ten independent factors (X1: age, X2: alcohol consumption, X3: hypertension, X4: bleeding, X5: blood transfusions, X6: general anesthesia, X7: intrathecal anesthesia, X8: D-dimer, X9: C-reactive protein, and X10: lymphocyte percentage) were identified as associated with an increased risk of VTE. Ten-fold cross-validation results showed that the ANN model was capable of predicting VTE in surgical patients, with an area under the curve (AUC) of 0.89, a Brier score of 0.01, an accuracy of 0.96, and a F1 score of 0.92. The ANN model slightly outperformed the logistic regression model and the Caprini model, but a DeLong test showed that the statistical difference in the AUCs of the ANN and logistic regression models was insignificant (P&gt;0.05). Conclusion: Ten statistical indicators relevant to VTE risk prediction for surgical patients were identified, and ANN and logistic regression both showed promising results as decision-supporting tools for VTE prediction.

Assessing the Impact of ATM and POS Transactions on Currency Circulation in Nigeria: A Comparison of Artificial Neural Network and Linear Regression Models

Assessing the Impact of ATM and POS Transactions on Currency Circulation in Nigeria: A Comparison of Artificial Neural Network and Linear Regression Models

Machine Learning-Assisted Determination of C6H14 Mole Fraction From Molecular Emissions of Laser-Induced Hexane-Air Plasmas.

Laser-induced plasmas of materials containing hydrocarbons present strong carbon molecular emission features. Using these emissions to build models relating changes in spectral features to a physical parameter of the system, such as hydrocarbon content, can be difficult because of the dynamic complexity of the spectral features and temperature disequilibrium between molecular species. This study presents machine learning models trained to quantify the mole fraction of hexane in hexane-air plasmas from CN Violet and C2 Swan spectral features. Ensemble regression methods provide the most accurate predictions with root mean squared error on the order 10-2. Artificial neural network regressions produce predictions with superlative sensitivity, exhibiting detection limits as low as 0.008. These foundational models can be further refined with more advanced data to quantify the presence of carbon species in complex plasma environments, such as high-speed reacting flows.

Prediction of unconfined compressive strength of lime treated soils

ABSTRACT Robust models based on Artificial Neural Network and simplified linear regression were proposed to predict the Unconfined Compressive Strength (UCS) of Lime Treated Soils (LTS). In total, an experimental database using 1120 test specimens was created. Critical examination of the collected experimental data suggested that there are eight key parameters that govern the attained strength gain. These input parameters are liquid limit, plastic limit, dry unit weight, water content, fine content, lime content, curing temperature and curing time whereas the only output dependent parameter is the UCS. The parameters of the proposed model including weights, biases and transfer functions were successfully converted to an explicit mathematical model relating the UCS with the key input parameters. Based on the results of the statistical evaluation, it was shown that a three-layered Artificial-Neural-Network model with 19 hidden neurons was capable to predict the UCS of (LTS) with a high degree of accuracy that was better than that achieved by the developed regression model. A coupling effect of the input parameters and weights analysis were conducted for the developed Artificial-Neural-Network-model to assess the importance of the key parameters. Analysis of weights of the model showed that curing time was the most significant factor affecting strength gain.