Prediction of the undrained shear strength of remolded soil with non-linear regression,fuzzy logic,and artificial neural network

This study aims to predict the undrained shear strength of remolded soil samples using non-linear regression analyses,fuzzy logic,and artificial neural network modeling.A total of 1306 undrained shear strength results from 230 different remolded soil test settings reported in 21 publications were collected,utilizing six different measurement devices.Although water content,plastic limit,and liquid limit were used as input parameters for fuzzy logic and artificial neural network modeling,liquidity index or water content ratio was considered as an input parameter for non-linear regression analyses.In non-linear regression analyses,12 different regression equations were derived for the prediction of undrained shear strength of remolded soil.Feed-Forward backpropagation and the TANSIG transfer function were used for artificial neural network modeling,while the Mamdani inference system was preferred with trapezoidal and triangular membership functions for fuzzy logic modeling.The experimental results of 914 tests were used for training of the artificial neural network models,196 for validation and 196 for testing.It was observed that the accuracy of the artificial neural network and fuzzy logic modeling was higher than that of the non-linear regression analyses.Furthermore,a simple and reliable regression equation was proposed for assessments of undrained shear strength values with higher coefficients of determination.

Pattern Analysis and Regressive Linear Measure for Botnet Detection

Capturing the distributed platform with remotely controlled compromised machines using botnet is extensively analyzed by various researchers.However,certain limitations need to be addressed efficiently.The provisioning of detection mechanism with learning approaches provides a better solution more broadly by saluting multi-objective constraints.The bots’patterns or features over the network have to be analyzed in both linear and non-linear manner.The linear and non-linear features are composed of high-level and low-level features.The collected features are maintained over the Bag of Features(BoF)where the most influencing features are collected and provided into the classifier model.Here,the linearity and non-linearity of the threat are evaluated with Support Vector Machine(SVM).Next,with the collected BoF,the redundant features are eliminated as it triggers overhead towards the predictor model.Finally,a novel Incoming data Redundancy Elimination-based learning model(RedE-L)is built to classify the network features to provide robustness towards BotNets detection.The simulation is carried out in MATLAB environment,and the evaluation of proposed RedE-L model is performed with various online accessible network traffic dataset(benchmark dataset).The proposed model intends to show better tradeoff compared to the existing approaches like conventional SVM,C4.5,RepTree and so on.Here,various metrics like Accuracy,detection rate,Mathews Correlation Coefficient(MCC),and some other statistical analysis are performed to show the proposed RedE-L model's reliability.The F1-measure is 99.98%,precision is 99.93%,Accuracy is 99.84%,TPR is 99.92%,TNR is 99.94%,FNR is 0.06 and FPR is 0.06 respectively.

Kinetics of thermal decomposition of lanthanum oxalate hydrate

Lanthanum oxalate hydrate La2（C2O4）3·10H2O,the precursor of La2O3 ultrafine powders,was prepared by impinging stream reactor method with PEG 20000 as surfactant.Thermal decomposition of La2（C2O4）3·10H2O from room temperature to 900 °C was investigated and intermediates and final solid products were characterized by FTIR and DSC-TG.Results show that the thermal decomposition process consists of five consecutive stage reactions.Flynn-Wall-Ozawa（FWO） and Kissinger-Akahira-Sunose（KAS） methods were implemented for the calculation of energy of activation（E）,and the results show that E depends on α,demonstrating that the decomposition reaction process of the lanthanum oxalate is of a complex kinetic mechanism.The most probable mechanistic function,G（α）=[1-（1＋α）1/3]2,and the kinetic parameters were obtained by multivariate non-linear regression analysis method.The average E-value that is compatible with the kinetic model is close to value which was obtained by FWO and KAS methods.The fitting curve matches the original TG curve very well.

Non-isothermal Decomposition Mechanism and Kinetics of LiClO_4 in Nitrogen

The non-isothermal decomposition kinetics of LiClO4 in flow N2 atmosphere was studied. TG-DTA curves show that the decomposition proceeded through two well-defined steps below 900℃, and the mass loss was in agreement with the theoretical value. XRD profile demonstrates that the product of the thermal decomposition at 500℃ is LiCI. For the decomposition kinetics study, the activation energies calculated with the Friedman method were considered as the initial values for non-linear regression and were used for verifying the correctness of the fired models. The decomposition process was fitted by a two-step consecutive reaction： extended Prout-Tompkins equation[Bna, f（α） is （1-α）^nα^α] followed by a lth order reaction（F1）. The activation energies were （215.6±0.2） and （251.6±3.6） kJ/mol, respectively. The exponentials n and a for Bna reaction were （0.25±0.05） and （0.795±0.005）, respectively. The reaction types and activation energies were in agreement with those obtained from the isothermal method, but the exponentials were optimized for better firing and prediction.

A regression model-based method for indoor positioning with compound location fingerprints

This paper proposed and evaluated an estimation method for indoor positioning.The method combines location fingerprinting and dead reckoning differently from the conventional combinations.It uses compound location fingerprints,which are composed of radio fingerprints at multiple points of time,that is,at multiple positions,and displacements between them estimated by dead reckoning.To avoid errors accumulated from dead reckoning,the method uses short-range dead reckoning.The method was evaluated using 16 Bluetooth beacons installed in a student room with the dimensions of 11×5 m with furniture inside.The Received Signal Strength Indicator(RSSI)values of the beacons were collected at 30 measuring points,which were points at the intersections on a 1×1 m grid with no obstacles.A compound location fingerprint is composed of RSSI vectors at two points and a displacement vector between them.Random Forests(RF)was used to build regression models to estimate positions from location fingerprints.The root mean square error of position estimation was 0.87 m using 16 Bluetooth beacons.This error is lower than that received with a single-point baseline model,where a feature vector is composed of only RSSI values at one location.The results suggest that the proposed method is effective for indoor positioning.

Mapping methods for output-based objective speech quality assessment using data mining

Objective speech quality is difficult to be measured without the input reference speech.Mapping methods using data mining are investigated and designed to improve the output-based speech quality assessment algorithm.The degraded speech is firstly separated into three classes(unvoiced,voiced and silence),and then the consistency measurement between the degraded speech signal and the pre-trained reference model for each class is calculated and mapped to an objective speech quality score using data mining.Fuzzy Gaussian mixture model(GMM)is used to generate the artificial reference model trained on perceptual linear predictive(PLP)features.The mean opinion score(MOS)mapping methods including multivariate non-linear regression(MNLR),fuzzy neural network(FNN)and support vector regression(SVR)are designed and compared with the standard ITU-T P.563 method.Experimental results show that the assessment methods with data mining perform better than ITU-T P.563.Moreover,FNN and SVR are more efficient than MNLR,and FNN performs best with 14.50% increase in the correlation coefficient and 32.76% decrease in the root-mean-square MOS error.

Intelligent process expert database of double pulse MIG welding of AI-Si alloy

Based on double pulse welding process characteristics, expert database structure and work flow are designed. Further, multiple outstanding specifications of 1.0 ram-diameter wire are obtained through a large number of experiments. By making non-linear regression analysis on these groups of standards, the relationship between average welding current and other pulse parameters can be found out. Polynomial regression equation is set up for further realization of＂ parameter estimation function of the expert database. Finally, the preliminary developed expert database is tested. The result indicates that the unified adjusting and parameters estimation of the expert database leads to stable welding process and good weld appearance.

Temperature change along elevation and its effect on the alpine timberline tree growth in the southeast of the Tibetan Plateau

Smith fir （Abies georgei var. smithii）, which is the timberline constructive tree species in the cool slope of Mt. Sygera in the southeast ofTibet, plays a very important role in maintaining the timberline completeness and indicating global climate change. This study uses theinstrumental recorded meteorological data along the altitude from 3600 to 4400 m at every 200 m in the growing season, investigates the smithfir growth biomass from 2006 to 2010 in the same timberline ecotone, and makes a non-linear regression analysis to determine the relationshipbetween the alpine tree growth biomass and its in-situ environment condition. The results showed that the cool and warm slope share different airtemperature lapse rates, which were
0.48 C （100 m）
1 in the warm slope and
0.54 C （100 m）
1 in the cool slope, respectively. However, thedominant timberline tree species in the warm slope was Sabina saltuaria, and it can reach as high as 4570 m, which is approximately 170 mhigher than that in the cool slope. Moreover, the smith fir in the cool slope was only distributed in the range of elevation from approximately3600 to 4400 m. The altitude of approximately 3800 m was the appropriate altitude for the growing smith fir, where the mean air temperature inthe growing season was about 9.0 C, and the young smith fir tree can form more biomass. The results suggested that alpine forest chose asuitable environment where trees can grow more in the prolonged succession, but not in the warmer or cooler condition, it could be seen as abiological evidence for climate change.

Mathematical model for abrasive suspension jet cutting based on orthogonal test design

This paper describes the application of orthogonal test design coupled with non-linear regression analysis to optimize abrasive suspension jet （AS J） cutting process and construct its cutting model. Orthogonal test design is applied to cutting stainless steel. Through range analysis on experiment results, the optimal process conditions for the cutting depth and the kerr ratio of the bottom width to the top width can be determined. In addition, the analysis of ranges and variances are all employed to identify various factors： traverse rate, working pressure, nozzle diameter, standoff distance which denote the importance order of the cutting parameters affecting cutting depth and the kerf ratio of the bottom width to the top width. ~rthermore, non-linear regression analysis is used to establish the mathematical models of the cutting parameters based on the cutting depth and the kerr ratio. Finally, the verification experiments of cutting parameters＇ effect on cutting performance, which show that optimized cutting parameters and cutting model can significantly improve the prediction of the cutting ability and quality of ASJ.

Support vector machine regression(SVR)-based nonlinear modeling of radiometric transforming relation for the coarse-resolution data-referenced relative radiometric normalization(RRN)

Radiometric normalization,as an essential step for multi-source and multi-temporal data processing,has received critical attention.Relative Radiometric Normalization(RRN)method has been primarily used for eliminating the radiometric inconsistency.The radiometric trans-forming relation between the subject image and the reference image is an essential aspect of RRN.Aimed at accurate radiometric transforming relation modeling,the learning-based nonlinear regression method,Support Vector machine Regression(SVR)is used for fitting the complicated radiometric transforming relation for the coarse-resolution data-referenced RRN.To evaluate the effectiveness of the proposed method,a series of experiments are performed,including two synthetic data experiments and one real data experiment.And the proposed method is compared with other methods that use linear regression,Artificial Neural Network(ANN)or Random Forest(RF)for radiometric transforming relation modeling.The results show that the proposed method performs well on fitting the radiometric transforming relation and could enhance the RRN performance.

Kinetics for Preparation of K_(2)Ti_(2)O_(5)Using TiO_(2)Microparticles and Nanoparticles as Precursors

The formation mechanism of K2Ti2O5 was investigated with Ti O2 microparticles and nanoparticles as precursors by the thermogravimetric(TG) technique. A method of direct multivariate non-linear regression was applied for simultaneous calculation of solid-state reaction kinetic parameters from TG curves. TG results show more regular decrease from initial reaction temperature with Ti O2 nanoparticles as raw material compared with Ti O2 microparticles, while mass losses finish at similar temperatures under the experimental conditions. From the mechanism and kinetic parameters, the reactions with the two materials are complex consecutive processes, and reaction rate constants increase with temperature and decrease with conversion. The reaction proceedings could be significantly hindered when the diffusion process of reactant species becomes rate-limiting in the later stage of reaction process. The reaction active sites on initial Ti O2 particles and formation of product layers may be responsible to the changes of reaction rate constant. The calculated results are in good agreement with experimental ones.

Mathematical approach for understanding deagglomeration behaviour of drug powder in formulations with coarse carrier

Deagglomeration of cohesive particles in combination with coarse carrier is a key requirement for inhaled formulations.The aim of the project was to propose a mathematical approach to understand aerosolization behaviour of micronized particles alone and in formulation with carriers.Salbutamol sulphate and salmeterol xinafoate were blended separately with fine lactose(ratio 1:4)and fine and coarse lactose(1:4:63.5).Laser diffraction was employed to characterize the powder median particle size.The deagglomeration of micronized materials followed an asymptotic monoexponential relationship.When the coarse lactose was added,the relationship fitted a bi-exponential equation showing an easily and a poorly dispersed fraction.Using model hydrophobic and hydrophilic APIs,this study has demonstrated the utility of an analytical approach that can parameterize deagglomeration behaviour of carrier-free and carrier-based inhalation formulations.The analytical approach provides the ability to systematically study the effect of material,formulation and processing factors on deagglomeration behaviour.

Application of deep autoencoder model for structural condition monitoring

Damage detection in structures is performed via vibra-tion based structural identification. Modal information, such as fre-quencies and mode shapes, are widely used for structural dama-ge detection to indicate the health conditions of civil structures.The deep learning algorithm that works on a multiple layer neuralnetwork model termed as deep autoencoder is proposed to learnthe relationship between the modal information and structural stiff-ness parameters. This is achieved via dimension reduction of themodal information feature and a non-linear regression against thestructural stiffness parameters. Numerical tests on a symmetri-cal steel frame model are conducted to generate the data for thetraining and validation, and to demonstrate the efficiency of theproposed approach for vibration based structural damage detec-tion.

Adaptability and Stability Analysis of Soybean Genotypes Using Toler and Centroid Methods

Soybean (Glycine max L. Merr.) adaptation to new environments has been hard to predict based on maturity group. The aim of this study was to evaluate the performance of 14 soybean genotypes, from the Soybean Breeding Program of the Federal University of Uberlandia, in their adaptive capacity and seed yield stability at 3 locations and 2 growing seasons. For the adaptability and stability analysis the Toler and Centroid methods were used;5 genotypic groups were identified in the first whereas 4 groups were identified in the latter. By the Toler method group A was composed by 4 genotypes, UFU-001, UFU-003, UFU-0010, and UFU-001. They showed a convex pattern of adaptability and stability. In contrast, the genotypes UFU-008 and UFU-0013 were classified in Group E with a concave pattern of adaptability and stability. Regarding results from the Centroid method, the Genotype UFU-002, with higher seed yield than average, was the only genotype in Ideotype VI with moderate adaptability to favorable environments. In contrast, 10 genotypes were included in the Ideotype V, of medium general adaptability. The genotypes UFU-001, UFU-002, UFU-006, UFU-0010, and UFU-0011 were recommended for use in the Brazilian Cerrado growing region. These genotypes had high seed yield potential in high quality environments.

Application of retardation time spectra of asphalt

To improve the performance of asphalt pavement, the dynamic and static tests of asphalt were used to measure its viscoelastic properties under different time. Based on the obtained data of static creep compliances and dynamic compliances according to the static creep test and dynamic test of asphalt using the dynamic shear rheometer, the discrete retardation time spectra were attained using the non-linear regression method. All viscoelastic functions are mathematically equivalent and belong to the same retardation time spectra, so the dynamic compliances of asphalt were converted to the static creep compliance using the retardation time spectra. Good correlations were found between calculation results and measurement results. In accordance to these findings, the retardation time spectra can accurately transform static and dynamic viscoelastic functions. Therefore, we can obtain viscoelastic properties over much larger time or frequency region than measurement results.

Modeling of the Global Daily Horizontal Solar Radiation Data over Togo

Solar photovoltaic appears to be the most interesting renewable energy in developing countries where its deposit is abundant. Unfortunately, the lack of precise knowledge of solar radiation deposit and its limited data hinder optimal exploitation of solar installations. This study presents a performing model for daily global horizontal solar radiation for the five regional capitals in Togo: Lomé, Atakpamé, Sokodé, Kara and Dapaong. The data used for the study were obtained from the General Directorate of National Meteorology of Togo, for five years. The model developed combines linear and nonlinear methods with harmonic and exponential terms taking into account climatological parameters such as location latitude, daily relative humidity, daily ratio of sunshine duration and daily mean temperature. Statistical errors of the model were compared to those of two previous models elaborated for Togo and Nigeria. The results showed that the model is more efficient to predict global horizontal solar radiation over the five main cities in Togo. The comparison of estimated data and measured ones showed a good agreement between them.

Kinetic, Isotherm and Thermodynamic Studies of the Adsorption of Thymol Blue onto Powdered Activated Carbons from <i>Garcinia cola Nut</i>Shells Impregnated with H₃PO₄and KOH: Non-Linear Regression Analysis

In this work, activated carbons (ACs) prepared by chemical activation of garcinia cola nut shell impregnated with H3PO4 (CBH2/1) and KOH (CBK1/1) were used to study the kinetics, equilibrium and thermodynamics of the adsorption of thymol blue from aqueous solution. The characterization of ACs showed the BET measurements gave surface area and total pore volume respectively of 328.407 m2·g-1 and 0.1032 cm3·g-1 for CBH2/1 and 25.962 m2·g-1 and 0.03 cm3·g-1for CBK1/1;elemental analysis showed a high percentage of carbon in both ACs. Influence of parameters such as initial pH, contact time, adsorbent mass, initial concentration, ionic strength and the effect of temperature on the removal of thymol blue from aqueous solution were studied in batch mode. The studies showed that equilibrium adsorption was attained after 60 minutes for the two ACs, adsorption capacity increased with increasing concentration of thymol blue, and maximum adsorption capacity was obtained at an acidic environment with pH 2. Avrami’s non-linear kinetic expression was the best suited for describing the adsorption kinetics of thymol blue onto ACs, while equilibrium data showed that the three-parameter isotherms better described the adsorption process since R2 > 0.96, and the error functions were lowest for all of them. Maximum adsorption capacity values obtained using the three-parameter Fritz-Schlunder equation were 32.147 mg·g-1 for CBH2/1 and 67.494 mg·g-1 for CBK1/1. The values of the model parameters g and mFS respectively, obtained using the Redlich-Peterson and Fritz-Schlunder III isotherms below 1, showed that the adsorption of thymol blue by the ACs occurred on heterogeneous surfaces. Thermodynamic analyses of the data of the adsorption of thymol blue onto ACs revealed that the adsorption process was temperature dependent, endothermic and spontaneous.

Nitrogen Release from Slow-Release Fertilizers in Soils with Different Microbial Activities

Soil microbial activity is recognized as an important factor affecting nitrogen (N) release from slow-release fertilizers. However,studies on the effect of size and activity of soil microflora on fertilizer degradation have provided contrasting results. To date, no clear relationships exist between soil microbial activity and the release of N from slow-release fertilizers. Hence, the aim of this study was to better understand such relationships by determining the release of N from three slow-release fertilizers in soils with different microbial activities. Soils were amended with urea-formaldehyde (UF), isobutylidene diurea (IBDU), and crotonylidene diurea (CDU). Urea, a soluble fertilizer, was used as the control. Fertilized soil samples were placed in a leaching system, and the release of N was determined by measuring ammonium-N and nitrate-N concentrations in leachates during 90 d of incubation. Non-linear regression was used to fit N leaching rate to a first-order model. In all the treated soils, N was released in the order: urea (89%–100%) > IBDU (59%–94%) >UF (46%–73%) > CDU (44%–56%). At the end of incubation, N released from CDU did not differ (P > 0.05) among soils. On the contrary, UF and IBDU released significantly lower (P < 0.05) amounts of N in the soil with higher microbial activity and lower pH.The rate constant (K_0) for UF was lower (P < 0.05) in the soil with lower pH. Taken together, our results indicated that soil microbial size and microbial activity had a marginal effect on fertilizer mineralization.

Machine learning prediction of magnetic properties of Fe-based metallic glasses considering glass forming ability

Fe-based metallic glasses(MGs)have shown great commercial values due to their excellent soft magnetic properties.Magnetism prediction with consideration of glass forming ability(GFA)is of great signifi-cance for developing novel functional Fe-based MGs.However,theories or models established based on condensed matter physics exhibit limited accuracy and some exceptions.In this work,based on 618 Fe-based MGs samples collected from published works,machine learning(ML)models were well trained to predict saturated magnetization(B_(s))of Fe-based MGs.GFA was treated as a feature using the experimental data of the supercooled liquid region(△T_(x)).Three ML algorithms,namely eXtreme gradient boosting(XGBoost),artificial neural networks(ANN)and random forest(RF),were studied.Through feature selection and hyperparameter tuning,XGBoost showed the best predictive performance on the randomly split test dataset with determination coefficient(R^(2))of 0.942,mean absolute percent error(MAPE)of 5.563%,and root mean squared error(RMSE)of 0.078 T.A variety of feature importance rankings derived by XGBoost models showed that T_(x) played an important role in the predictive performance of the models.This work showed the proposed ML method can simultaneously aggregate GFA and other features in ther-modynamics,kinetics and structures to predict the magnetic properties of Fe-based MGs with excellent accuracy.

Effects of dietary protein levels on the long-term growth response and fitting growth models of gibel carp(Carassius auratus gibelio)

A 41-wk growth trial was conducted to evaluate the effects of dietary protein levels on the long-term growth response and fitting growth models of gibel carp(Carassius auratus gibelio) with an initial body weight of 1.85 ± 0.17 g. The dietary protein levels were designed at 320(P32), 360(P36). 400(P40).and 440 g/kg(P44), respectively. The growth curves of the gibel carp for each group were fitted and analyzed with four nonlinear regression models(Gompertz. logistic. von Bertalanffy and Richards). The final body weights(mean ± SD) of the fish were 226 ± 6.231 ± 7.242 ± 2, and 236 ± 2 g for P32, P36, P40,and P44. respectively. Feed conversion ratio of P40 and P44 groups was significantly lower than that of P32 and P36 groups(P < 0.05). Productive protein value of P44 group was significantly lower than that of P32 and P36 groups, but not different from that of P40 group(P > 0.05). The growth response of the gibel carp for each group was the best fitted by Richards model with the lowest Chi^2, residual sum of squares and residual variance, then Gompertz and von Bertalanffy growth models, but the logistic model did not fit the data well justified by Chi^2 values. The optimal protein level(400 g/kg) prolonged the stage of fast growth and predicted the highest asymptotic weight, which was close to the harvest size in practice.