Pace Regression Research Articles

A novel additive regression model for streamflow forecasting in German rivers

Forecasting streamflows, essential for flood mitigation and the efficient management of water resources for drinking, agriculture and hydroelectric power generation, presents a formidable challenge in most real-world scenarios. In this study, two models, the first based on the Additive Regression of Radial Basis Function Neural Networks (AR-RBF) and the second based on the stacking with the Pace Regression of the Multilayer Perceptron and Random Forest (MLP-RF-PR), were compared for the prediction of short-term (1–3 days ahead) and medium-term (7 days ahead) daily streamflow rates of three different rivers in Germany: the Elbe River at Wittenberge, the Leine River at Herrenhausen, and the Saale River at Hof. The lagged values of streamflow rate, precipitation and temperature were considered for the modeling. Moreover, the Bayesian Optimization (BO) algorithm was used to assess the optimal number of lagged values and hyperparameters. Both models showed accurate predictions for short-term forecasting, with R2 for 1-day ahead predictions ranging from 0.939 to 0.998 for AR-RBF and from 0.930 to 0.996 for MLP-RF-PR, while MAPE ranged from 2.02 % to 8.99 % for AR-RBF and from 2.14 % to 9.68 % for MLP-RF-PR, when exogeneous variables were included. As the forecast horizon increased, a reduction in forecasting accuracy was observed. However, both models could still predict the overall flow pattern, even for 7-day-ahead predictions, with R2 ranging from 0.772 to 0.871 for AR-RBF and from 0.703 to 0.840 for MLP-RF-PR, while MAPE ranged from 10.60 % to 20.45 % for AR-RBF and from 10.44 % to 19.65 % for MLP-RF-PR. Overall, the outcomes of this study suggest that both AR-RBF and MLP-RF-PR can be reliable tools for the short- and medium-term streamflow rate prediction, requiring a short number of parameters to be optimized, making them easy to implement while reducing the calculation time required.

Efficacy of machine learning algorithms in estimating emissions in a dual fuel compression ignition engine operating on hydrogen and diesel

Emission created by combustion of fossil fuels are a major concern of the world for the past few decades. The stringent emission norms have impacted the automobile manufacturers to work on exhaust emissions and its impact. This research focused on using machine learning regression models to evaluate the efficacy of experimental results for a dual fuel compression ignition (CI) engine operating on hydrogen and diesel. In the present study, engine emissions were estimated using 29 regression algorithms. A total of 5 input data namely, concentration of hydrogen, engine load, diesel intake, speed and equivalence ratio were considered in the study to estimate various emissions like oxides of nitrogen (NOx), carbon dioxide (CO2), hydrocarbon (HC) and smoke. Correlation coefficient, mean absolute error, root mean squared error, relative absolute error and root relative squared error were adopted as the performance metrics in the present study. Amongst the algorithms considered, pace regression, radial basis function regressor, multilayer perceptron regressor and alternating model tree produced the highest correlation coefficient of 0.9985, 0.8958, 0.9950 and 0.9256 in estimating the engine emissions like CO2, smoke, NOx and HC respectively. Additionally, an attempt was made to establish an individual algorithm that can estimate all the emissions was identified as multilayer perceptron regressor with correlation coefficient values of 0.9977 (CO2), 0.9950 (NOx), 0.8501(smoke) and 0.8731(HC) respectively.

Obtaining mathematical equations for energy, exergy and electrical efficiency: A machine learning approach

ABSTRACT In this study, in which the energy, exergy, and electrical efficiency values of the photovoltaic thermal panel are modeled with different machine learning algorithms, mathematical equations that can calculate the efficiency values have been obtained as an innovative approach. Data sets consisting of environmental parameters (temperature, wind speed, solar radiation, humidity) of the environment in which the experiments were carried out were used in the models. Thus, the effects of environmental parameters on collector efficiency values were observed, and mathematical equations were produced using these parameters with the help of the decision tree algorithm and Pace regression. In addition, environ-economic analyzes of the panels were made and the coefficient of performance values were examined. In the experiments, two data sets were obtained. With one of these data sets, the efficiency values were modeled with machine learning algorithms, and the accuracy of the mathematical equations obtained with the other data set was proven. The mean absolute percentage error values of the energy, exergy and electrical efficiency models created with the decision tree are 8.04%, 1.76%, and 1.43%, respectively. Similarly, Pace model error values are 3.83%, 2.54%, and 2.1%. The high accuracy values of the obtained efficiency equations under different experimental conditions show that these equations can be used under different conditions and in different solar energy systems.

Experimental Analysis of Different Refrigerants’ Thermal Behavior and Predicting Their Performance Parameters

This study experimentally compares thermodynamic performance of R290 and R404A refrigerants in a refrigeration system. In the first part of the paper, energy analysis of the refrigeration system was performed at various evaporator and condenser temperatures. Results revealed that R404A refrigerant caused an 18.6% increase in compressor power consumption. The highest coefficient of performance values in the system for R290 and R404A were 3.99 and 3.21, respectively. The second part of the paper includes artificial intelligence prediction studies. The pace and elastic net regression models were used to predict performance parameters. A single equation that can predict the cooling capacity and compressor power consumption of R290 and R404A simultaneously was derived. For the cooling capacity, pace regression showed mean absolute error of 0.0252 and root-mean-squared error of 0.0334, while elastic net regression indicated mean absolute error of 0.1103 and root-mean-squared error of 0.1262. It was concluded that R290 had better thermodynamic performance than R404A and the equations obtained with artificial intelligence were applicable to predict the experimental findings, regardless of which refrigerant gas was used.

Towards robust smart data-driven soil erodibility index prediction under different scenarios

Soil erosion is a major cause of damage to agricultural lands in many parts of the world and is of particular concern in semiarid parts of Iran. We use five machine learning techniques—Random Forest (RF), M5P, Reduced Error Pruning Tree (REPTree), Gaussian Processes (GP), and Pace Regression (PR)—under two scenarios to predict soil erodibility in the Dehgolan region, Kurdistan Province, Iran. Our models are based on a variety of soil properties, including soil texture, structure, permeability, bulk density, aggregates, organic matter, and chemical constituents. We checked the validity of the models with statistical metrics, including the coefficient of determination (R2), mean absolute error (MAE), root mean squared error (RMSE), T-tests, Taylor diagrams, and box plots. All five algorithms show a positive correlation between the soil erodibility factor (K) and silt, sand, fine sand, bulk density, and infiltration. The GP model has the highest prediction accuracy (R2 = 0.843, MAE = 0.0044, RMSE = 0.0050). It outperformed the RF (R2 = 0.812, MAE = 0.0050, RMSE = 0.0061), PR, (R2 = 0.794, MAE = 0.0037, RMSE = 0.0052), M5P (R2 = 0.781, MAE = 0.0043, RMSE = 0.0053), and REPTree (R2 = 0.752, MAE = 0.0045, RMSE = 0.0056) algorithms and thus is a useful complement to studies aimed at predicting soil erodibility in areas with similar climate and soil characteristics.

Stacking ensemble-based hybrid algorithms for discharge computation in sharp-crested labyrinth weirs

Labyrinth weirs are utilized to transport a greater discharge during floods in contrast to conventional weirs due to their increased weir crest length. Nevertheless, due to the increased geometric complexity of labyrinth weirs, determination of accurate discharge coefficients and accordingly, head-discharge ratings are quite essential issues in practical application. Hence, as a first step the present study proposes the following eight standalone algorithms: decision table (DT), Kstar, least median square (LMS), M5 prime (M5P), M5 rule (M5R), pace regression (PR), random forest (RF) and sequential minimal optimization (SMO). Then, applying the stacking (ST) algorithm, these standalone models were hybridized to predict the discharge coefficient (Cd) for sharp-crested labyrinth weirs. Potential/effective variables were constructed in the form of several independent dimensionless parameters (i.e., θ, h/W, L/B, L/h, Froude number (Fr), B/W and L/W) to predict Cd as an output. The accuracy of the developed models was examined in terms of different statistical visually based and quantitative-based error measurement criteria. The results illustrate that h/W and B/W parameters have the highest and lowest effect on the Cd prediction, respectively. According to NSE, all developed algorithms provided accurate performances, while ST-Kstar had the highest prediction power.

Investigation of the effect of deposit layer on heat transfer in the Triga Mark II nuclear research reactor cooling system

In this presented study, the cooling problem of the I.T.U. Triga Mark-II reactor has been handled and analyzed, and solutions were proposed. First of all, a thermal model of the reactor, heat exchanger, and cooling tower trio was established in the reactor. With this model, which was obtained with the help of experimental data, the parameters affecting the change of reactor water temperature over time were determined, and significant findings were obtained by investigating the possibilities of increasing the cooling power of the existing system. Then, using these mathematical equations, the effects of parameters that can affect the power of the reactor cooling system are investigated. The parameters affecting the cooling power are the cooling water flow rates in the second cooling circuits and the deposited layer that may exist as a result of numerical calculations. Different models have been created with machine learning algorithms (page regression, decision tree) to estimate the effect of the deposit layer. The mathematical and predictive models obtained with the experimental data for the heat transfer coefficient of the deposit layer, hbd, were compared. The pace regression algorithm modeled the hbd values with the least error rate (RMSE: 1.66) among the models. It has been calculated that the average tank water temperature will decrease by approximately 3.5?C if the deposits layer is cleared.

Iterative classifier optimizer-based pace regression and random forest hybrid models for suspended sediment load prediction.

Suspended sediment load is a substantial portion of the total sediment load in rivers and plays a vital role in determination of the service life of the downstream dam. To this end, estimation models are needed to compute suspended sediment load in rivers. The application of artificial intelligence (AI) techniques has become popular in water resources engineering for solving complex problems such as sediment transport modeling. In this study, novel integrative intelligence models coupled with iterative classifier optimizer (ICO) are proposed to compute suspended sediment load in Simga station in Seonath river basin, Chhattisgarh State, India. The proposed models are hybridization of the random forest (RF) and pace regression (PR) models with the iterative classifier optimizer (ICO) algorithm to develop ICO-RF and ICO-PR hybrid models. The recommended models are established using the discharge and sediment daily data spanning a 35-year period (1980-2015). The accuracy of the developed models is examined in terms of error; by root mean square error (RMSE) and mean absolute error (MAE); and based on a correlation index of determination coefficient (R2). The proposed novel hybrid models of ICO-RF and ICO-PR have been found to be more precise than their stand-alone counterparts of RF and PR. Overall, ICO-RF models delivered better accuracy than their alternatives. The results of this analysis tend to claim the appropriateness of the implemented methodology for precise modeling of the suspended sediment load in rivers.

Antimicrobial Isoflavones and Derivatives from Erythrina (Fabaceae): Structure Activity Perspective (Sar & Qsar) on Experimental and Mined Values Against Staphylococcus aureus

Prenylated (iso)flavonoids, -flavans and pterocarpans from taxa in Erythrina are repeatedly flagged as potent antimicrobial compounds. In the current study, bark from E. lysistemon was extracted and seven isoflavone derivatives were purified: erybraedin A (1), phaseollidin (2), abyssinone V-4′ methyl ether (3), eryzerin C (4), alpumisoflavone (5), cristacarpin (6) and lysisteisoflavone (7). Minimum inhibition concentration (MIC) values were determined against a range of species of bacteria (skin pathogens), then values for another 67 derivatives from Erythrina, only against Staphylococcus aureus, were mined from the literature. Of the seven isolates, MIC values widely ranged from 1–600 μg/mL, with no obvious pattern of selectivity for Gram-types. Nevertheless, using the mined and experimentally determined values against S. aureus, Klekota-Roth fragments (Structure Activity Relationship: SAR) were determined then used as molecular descriptors to make a ‘decision tree’ based on structural characters inspired by the classes of antimicrobial potency (classes A-D). Furthermore, to make quantitative predictions of MIC values (Quantitative SAR: QSAR) ‘pace regression’ was utilized and validated (R² = 0.778, Q² = 0.727 and P² = 0.555). Evidently, the position and degree of prenylation is important; however, the presence of hydroxyl groups at positions 5 and 7 in ring A and 4′ in ring B is associated with lower MIC values. While antimicrobial results continue to validate the traditional use of E. lysistemon extracts (or Erythrina generally) in therapeutic applications consistent with anti-infection, it is surprising that this class of compound is not being utilized more often in general industry applications, such as food or cosmetic preservation, or in topical antimicrobial creams. Prenylated (iso)flavonoids are derived from several other Genera, such as Dorstenia (Moraceae), Ficus (Moraceae), Glycyrrhiza (Fabaceae), Paulownia (Lamiales) or Pomifera (Moraceae).

Fusion of PACE Regression and Decision Tree for Comment Volume Prediction

Fusion of PACE Regression and Decision Tree for Comment Volume Prediction

BLITE-SVR: New forecasting model for late blight on potato using support-vector regression

Various simple statistical methods have been used for the prediction of plant-disease epidemics. However, the need to develop a new model, reflecting many changed environmental factors and applicable to the Korean domestic farmhouse, has been raised. Given this point, we developed the potato late blight prediction model called BLITE-SVR, after which we predicted and verified the first date of occurrence with the data from 1976 to 1985 and from 2009 to 2012 through support-vector regression (SVR), a statistical method offering good performance. For the prediction model, we collected 13 kinds of weather data, including temperature, humidity, evaporation, and so on, which displayed very high correlation to the first date of the occurrence of late blight. The performance of BLITE-SVR has been evaluated through comparison with the conventional moving-average method that was previously used, as well as through pace regression and linear regression. The accuracy of prediction for the first date of occurrence was 64.3% by BLITE-SVR, thus showing a higher degree of accuracy compared with 42.9% by the conventional moving-average method, 42.9% by pace regression and 35.7% by linear regression. This study will enable farmers to match the targeted fungicide application to the time of greatest need and thereby achieve a reduction in chemical use.

Comparing single- and two-segment statistical models with a conceptual rainfall-runoff model for river streamflow prediction during typhoons

This study examined various regression-based techniques and an artificial neural network used for streamflow forecasting during typhoons. A flow hydrograph was decomposed into two segments, rising and falling limbs, and the individual segments were modeled using statistical techniques. In addition, a conceptual rainfall–runoff model, namely the Public Works Research Institute (PWRI)-distributed hydrological model, and statistical models were compared. The study area was the Tsengwen Reservoir watershed in Southern Taiwan. The data used in this study comprised the observed watershed rainfalls, reservoir inflows, typhoon characteristics, and ground weather data. The forecast horizons ranged from 1 to 12 h. A series of assessments, including statistical analyses and simulations, was conducted. According to the improvements in errors, among single-segment statistical models, the multilayer perceptron achieved superior prediction accurary compared with the regression-based methods. However, the pace regression was the most favorable according to an evaluation of model complexity and accuracy. To examine the robustness of the results for forecast horizons varying from 1 to 12 h, statistical significance tests were performed for the single- and two-segment models. The prediction ability of the two-segment models was superior to that of the single-segment models. In addition, Typhoon Sinlaku in 2008 was considered in a comparison between the conceptual PWRI model output and that of the developed statistical models. The results showed that the PWRI model yielded the least favorable results.

Determining a new formula for calculating low-density lipoprotein cholesterol: data mining approach.

Low-density lipoprotein cholesterol (LDL-C) is a risk factor of coronary heart diseases. The estimation of LDL-C (LDL-Cal) level was performed using Friedewald's equation for triglyceride (TG) level less than 400 mg/dL. Therefore, the aim of this study is to generate a new formula for LDL-Cal and validate the correlation coefficient between LDL-Cal and LDL-C directly measured (LDL-Direct). A data set of 1786 individuals receiving annual medical check-ups from the Faculty of Medical Technology, Mahidol University, Thailand in 2008 was used in this study. Lipid profiles including total cholesterol (TC), TG, high-density lipoprotein cholesterol (HDL-C) and LDL-C were determined using Roche/Hitachi modular system analyzer. The estimated LDL-C was obtained using Friedewald's equation and the homogenous enzymatic method. The level of TG was divided into 6 groups (TG<200, <300, <400, <500, <600 and < 1000 mg/dL) for constructing the LDL-Cal formula. The pace regression model was used to construct the candidate formula for the LDL-Cal and determine the correlation coefficient (r) with the LDL-Direct. The candidate LDL-Cal formula was generated for 6 groups of TG levels that displayed well correlation between LDL-Cal and LDL-Direct. Interestingly, The TG level was less than 1000 mg/dL, the regression model was able to generate the equation as shown as strong r of 0.9769 with LDL-Direct. Furthermore, external data set (n = 666) with TG measurement (36-1480 mg/dL) was used to validate new formula which displayed high r of 0.971 between LDL-Cal and LDL-direct. This study explored a new formula for LDL-Cal which exhibited higher r of 0.9769 and far beyond the limitation of TG more than 1000 mg/dL and potential used for estimating LDL-C in routine clinical laboratories.

Logistic-weighted regression improves decoding of finger flexion from electrocorticographic signals.

One of the most interesting applications of brain computer interfaces (BCIs) is movement prediction. With the development of invasive recording techniques and decoding algorithms in the past ten years, many single neuron-based and electrocorticography (ECoG)-based studies have been able to decode trajectories of limb movements. As the output variables are continuous in these studies, a regression model is commonly used. However, the decoding of limb movements is not a pure regression problem, because the trajectories can be apparently classified into a motion state and a resting state, which result in a binary property overlooked by previous studies. In this paper, we propose an algorithm called logistic-weighted regression to make use of the property, and apply the algorithm to a BCI system decoding flexion of human fingers from ECoG signals. Our results show that the application of logistic-weighted regression improves decoding performance compared to the application of linear regression or pace regression. The proposed algorithm is also immensely valuable in the other BCIs decoding continuous movements.

A comparison of different regression models for downscaling Landsat and MODIS land surface temperature images over heterogeneous landscape

Remotely sensed high spatial resolution thermal images are required for various applications in natural resource management. At present, availability of high spatial resolution (<200m) thermal images are limited. The temporal resolution of such images is also low. Whereas, coarser spatial resolution (∼1000m) thermal images with high revisiting capability (∼1day) are freely available. To bridge this gap, present study attempts to downscale coarser spatial resolution thermal image to finer spatial resolution using relationships between land surface temperature (LST) and vegetation indices over a heterogeneous landscape of India. Five regression based models namely (i) Disaggregation of Radiometric Temperature (DisTrad), (ii) Temperature Sharpening (TsHARP), (iii) TsHARP with local variant, (iv) Least median square regression downscaling (LMSDS) and (v) Pace regression downscaling (PRDS) are applied to downscale LST of Landsat Thematic Mapper (TM) and Terra MODIS (Moderate Resolution Imaging Spectroradiometer) images. All the five models are first evaluated on Landsat image aggregated to 960m resolution and downscaled to 480m and 240m resolution. The downscale accuracy is achieved using LMSDS and PRDS models at 240m resolution at 0.61°C and 0.75°C respectively. MODIS data downscaled from 1000m to 250m spatial resolution results root mean square error (RMSE) of 1.43°C and 1.62°C for LMSDS and PRDS models, respectively. The LMSDS model is less sensitive to outliers in heterogeneous landscape and provides higher accuracy when compared to other models. Downscaling model is found to be suitable for agricultural and vegetated landscapes up to a spatial resolution of 250m but not applicable to water bodies, dry river bed sand sandy open areas.

A qualitative and quantitative study of color emotion using valence-arousal

This paper describes qualitative and quantitative analysis of color emotion dimension expression using a standard device-independent colorimetric system. To collect color emotion data, 20 subjects are required to report their emotion responses, using a valence-arousal emotion model, to 52 color samples that are chosen from CIELAB Lch color spaces. Qualitative analysis, including analysis of variance (ANOVA), Pearson’s correlation and Spearman’s rank correlation, shows that significant differences exist between responses to achromatic and chromatic stimuli, but there are no significant differences between chromatic samples. There is a positive linear relationship between lightness/chroma and valence-arousal dimensions. Subsequently, several classic predictors are used to quantitatively predict emotion induced by color attributes. Furthermore, several explicit color emotion models are developed by using multiple linear regression with stepwise and pace regression. Experimental results show that chroma and lightness have stronger effects on emotions than hue, which is consistent with our qualitative results and other psychological studies. Arousal has greater predictive value than valence.

Predicting Number of Zombies in DDoS Attacks Using Pace Regression Model

Normal 0 false false false A DDoS attacker attempts to disrupt a target, by flooding it with illegitimate packets which are generated from a large number of zombies, usurping its bandwidth and overtaxing it to prevent legitimate inquiries from getting through. This paper reports the evaluation results of proposed approach that is used to predict number of zombies using Pace Regression Model. A relationship is established between number of zombies and observed deviation in sample entropy. Various statistical performance measures, such as R2, CC, SSE, MSE, RMSE, NMSE, η, MAE are used to measure the performance of the regression model. Network topologies similar to Internet used for simulation are generated using Transit-Stub model of GT-ITM topology generator. NS-2 network simulator on Linux platform is used as simulation test bed for launching DDoS attacks with varied number of zombies. The simulation results are promising as we are able to predict number of zombies efficiently using Pace Regression Model with considerably less error rate.

Modeling of mechanical properties and bond relationship using data mining process

Reinforced concrete is a widely used construction material. Its properties depend on the bond between the reinforcing bar and concrete as much as the compressive strength or properties of the reinforcing bar because of component of construction expose to both flexural and bond together compressive loads. In this paper, the bond properties of concretes with different mix designs were investigated according to the results of compressive, flexural, bond, and flexural-bond tests. The data mining (DM) process was used to determine relationships among the test results and DM algorithms. Seventeen modeling techniques within WEKA were applied to the experimental data for the prediction of bond properties. The results show that the implemented models were good at predicting the bond properties. The best results were obtained from the RepTree algorithm for bond strength, the Multilayer Perceptron algorithm for flexural-bond strength, the MedSq algorithm for bond slippage, and the Pace Regression for flexural-bond deformation. Bond and flexural-bond can be easily predicted using the compressive strength, flexural strength and tensile stress of the rebar. Although a relationship is also existent between these and bond slippage and flexural-bond deformation, these relationships are weaker than the others. These results suggested that the DM algorithms can be used as an alternative approach to predict the bond strength using the results of compressive, flexural, bond, and flexural-bond tests as input parameters. The predictions of the bond slippage and flexural-bond deformation models poorly agreed with the experimental results. It can be obtained more successful results for these variables, when DM models with different inputs like the rebar-concrete interface stress together the measured parameters are used.

A Study on GPS GDOP Approximation Using Support-Vector Machines

Global Positioning System (GPS) has extensively been used in various fields. Geometric Dilution of Precision (GDOP) is an indicator showing how well the constellation of GPS satellites is geometrically organized. GPS positioning with a low GDOP value usually gains better accuracy. However, the calculation of GDOP is a time- and power-consuming task that involves complicated transformation and inversion of measurement matrices. When selecting from many GPS constellations the one with the lowest GDOP for positioning, methods that can fast and accurately obtain GPS GDOP are imperative. Previous studies have shown that numerical regression on GPS GDOP can get satisfactory results and save many calculation steps. This paper deals with the approximation of GPS GDOP using statistics and machine learning methods. The technique of support vector machines (SVMs) is mainly focused. This study compares the performance of several methods, such as linear regression, pace regression, isotonic regression, SVM, artificial neural networks, and genetic programming (GP). The experimental results show that SVM and GP have better performance than others.

The prediction of the electric field level in the reverberation chamber depending on position of stirrer

This study presents nine modeling techniques within data mining process for the prediction of electric field level which depends on the positions of stirrer in the mode stirred reverberation chamber. These are Linear Regression (LR), Multi Layer Perceptron (MLP), Pace Regression (PR), KStar, Regression by discretization (RD), Random SubSpace (RS), M5P, RepTree and Decision Table models. Relations depending on frequency and position of stirrer have been carried out for the determination of electric field in the reverberation chamber. Obtained model results for each stirrer steps have been compared and the best model has been investigated. Results indicate that the use of derived formulation from these techniques will facilitate design and optimize of location of maxima (mode) and minima (null) of electric field in the reverberation chamber which a component of especially mode stirrer reverberation chamber test method.