Decision Tree Regression Algorithm Research Articles

Fast dentification of overlapping fluorescence spectra of oil species based on LDA and two-dimensional convolutional neural network

Although most petroleum oil species can be identified by their fluorescence spectra, overlapping fluorescence spectra make identification difficult. This study aims to address the issue that fluorescence spectroscopy is ineffective in identifying overlapping oil species. In this study, an equivalent model of overlapping oil species with fluorescence spectra was established. The linear discriminant analysis (LDA)-assisted machine learning (ML) algorithms K nearest neighbor (KNN), decision tree (DT), and random forest (RF) improved the identification of fluorescent spectrally overlapping oil species for diesel-lubricant oils. The identification accuracies of two-dimensional convolutional neural network (2DCNN), LDA combined with the ML algorithms effectively all 100 %. Furthermore, Partial Least Squares Regression (PLSR) algorithm, Support Vector Regression (SVR) algorithm, DT regression algorithm, and RF regression algorithm were also used to identify the lubricant concentration in diesel-lubricant oils. The coefficient of determination of the DT was 1, and the root-mean-square error was 0, which identified the concentration of lubricant oils in them accurately and without error.

Echocardiography Versus Magnetic Resonance Imaging Quantification and Novel Algorithm for Isolated Severe Tricuspid Regurgitation

Transthoracic echocardiography (TTE) is the first-line tool to evaluate isolated tricuspid regurgitation (TR) but it has limitations and its TR quantification compared with magnetic resonance imaging (MRI) has been studied infrequently. We compared isolated severe TR quantification by TTE against MRI and developed a novel TTE-based algorithm. Isolated TR patients graded severe by TTE and who underwent MRI January 2007 to June 2019 were studied. The TTE and MRI measurements were analyzed by correlation, area under receiver-operative characteristics curve (AUC), and classification and regression tree algorithm of TTE parameters to best identify MRI-derived severe TR (regurgitant volume ≥45ml and/or fraction ≥50%). A total of 108 of 262 (41%) that were graded as severe TR by TTE also had severe TR by MRI. There were moderate correlations between TTE and MRI in the quantification of TR severity and right atrial size (Pearson r=0.428 to 0.645) but none to modest correlations between them in right ventricle quantification. The key TTE parameters to identify MRI-derived severe TR in the decision tree regression algorithm were right atrial volume indexed ≥47ml/m2 and effective regurgitant orifice area ≥0.45cm2 and especially if there is right ventricle free wall strain ≥ -9.5%. This novel algorithm has an AUC of 0.76% and 79% agreement to detect severe TR by MRI, which higher than the American Society of Echocardiography criteria with AUC 0.68% and 66% agreement (p=0.006 and p <0.001, respectively). In conclusion, TTE-derived TR and right atrial quantification had moderate correlation and discrimination of severe TR by MRI, from which a novel TTE algorithm was derived, which had incrementally a higher accuracy than contemporary guidelines' criteria alone.

Development of a machine learning regression model for accurate sugarcane crop yield prediction, Jinja – Uganda

Sugarcane is one of the key crops grown worldwide and used for sugar processing, food, alcohol, biogas, fertilizer, and other products. There is a problem with Sugarcane yield prediction, yields aren’t accurately predicted, and this creates an impact on yields. This research looks at identifying methods used for the prediction, design, development, and evaluation of the three machine learning regression models used for predicting sugarcane yields in Uganda. This research was implemented using Data Science methodology, several machine learning algorithms for prediction of yields on dataset have been analyzed. The collected and analyzed dataset in this research had one output/ dependent variable and eight independent variables. The algorithms used to develop the prediction models are the Multiple Linear Regression algorithm, Decision Tree Regression algorithm, and Random Forest Regression algorithm to predict the output. The dataset of 3 years, 2019, 2020, and 2022 was considered and merged to train and test the model at a ratio of 80% to 20%. The accuracies of the individual models were compared after training, testing the dataset, and evaluation. The multiple Linear regression model results indicate that out of 100%, the model accuracy was 76.5%, the Decision Tree Regression Model scored 89.2%, Random Forest Regression Model was 94.6%. The random forest model came out as the best model. The Random Forest model has a percentage improvement of 60.4%. In future research, researchers can work on, A web-based machine learning model, Deep learning methods used to improve the model and more data can be used to improve the accuracy

Unimolecular dissociation of C6H6-C6H5Cl, C6H6-C6H3Cl3, and C6H6-C6Cl6 complexes using machine learning approach.

The application of Machine Learning (ML) algorithms in chemical sciences, particularly computational chemistry, is a vastly emerging area of modern research. While many applications of ML techniques have already been in place to use ML based potential energies in various dynamical simulation studies, specific applications are also being successfully tested. In this work, the ML algorithms are tested to calculate the unimolecular dissociation time of benzene-hexachlorobenzene, benzene-trichlorobenzene, and benzene-monochlorobenzene complexes. Three ML algorithms, namely, Decision-Tree-Regression (DTR), Multi-Layer Perceptron, and Support Vector Regression are considered. The algorithms are trained with simulated dissociation times as functions (attributes) of complexes' intramolecular and intermolecular vibrational energies. The simulation data are used for an excitation temperature of 1500K. Considering that the converged result is obtained with 1500 trajectories, an ML algorithm trained with 700 simulation points provides the same dissociation rate constant within statistical uncertainty as obtained from the converged 1500 trajectory result. The DTR algorithm is also used to predict 1000K simulation results using 1500K simulation data.

Novel echocardiography-based algorithms to identify severe aortic or mitral regurgitation by cardiac magnetic resonance imaging: a prospective multi-modality imaging study

Abstract Background Magnetic resonance imaging (MRI) has shown incremental prognostic value to transthoracic echocardiography (TTE) for valve diseases, however TTE remains more widely available and utilized. We sought to develop TTE-based algorithms that best identify severe AR and MR by MRI in a prospective study. Methods Patients with &amp;gt;moderate-to-severe AR (n=101) and MR (n=71) undergoing both TTE and CMR within 3-months (median 1 day) were prospectively studied. Correlation analyses were performed, and TTE-based decision tree regression algorithms were derived to best identify CMR-defined severe AR and MR (regurgitant fraction &amp;gt;40% and &amp;gt;33% respectively). Results Mean regurgitant volumes/fractions by TTE were 15 mL/18% and 13mL/18% higher than CMR in AR and MR respectively. Decision-tree analyses found regurgitant volume &amp;gt;50 mL and left ventricular end-systolic volume indexed &amp;gt;38 mL/m2 by TTE to best identify CMR-derived severe AR, and regurgitant fraction &amp;gt;50% and left ventricular stroke volume indexed &amp;gt;39 mL/m2 by TTE to best identify CMR-derived severe MR (Figure 1). Their areas under curve compared with current guidelines criteria were 0.75 versus 0.63 (P=0.038) for severe AR, and 0.80 versus 0.67 (P=0.017) for severe MR. Conclusion Novel TTE-based algorithms were devised to identify severe MRI-derived AR and MR, superior to the more complex multi-parametric criteria of current TTE guidelines. Funding Acknowledgement Type of funding sources: None.

Machine Learning Analysis on the Performance of Dye-Sensitized Solar Cell—Thermoelectric Generator Hybrid System

In cases where a dye-sensitized solar cell (DSSC) is exposed to light, thermal energy accumulates inside the device, reducing the maximum power output. Utilizing this energy via the Seebeck effect can convert thermal energy into electrical current. Similar systems have been designed and built by other researchers, but associated tests were undertaken in laboratory environments using simulated sunlight and not outdoor conditions with methods that belong to conventional data analysis and simulation methods. In this study four machine learning techniques were analyzed: decision tree regression (DTR), random forest regression (RFR), K-nearest neighbors regression (K-NNR), and artificial neural network (ANN). DTR algorithm has the least errors and the most R2, indicating it as the most accurate method. The DSSC-TEG hybrid system was extrapolated based on the results of the DTR and taking the worst-case scenario (node-6). The main question is how many thermoelectric generators (TEGs) are needed for an inverter to operate a hydraulic pump to circulate water, and how much area is required for that number of TEGs. Considering the average value of the electric voltage of the TEG belonging to node-6, 60,741 pieces of TEGs would be needed, which means about 98 m2 to circulate water.

A machine learning-based on-demand sweat glucose reporting platform

Diabetes is a chronic endocrine disease that occurs due to an imbalance in glucose levels and altering carbohydrate metabolism. It is a leading cause of morbidity, resulting in a reduced quality of life even in developed societies, primarily affected by a sedentary lifestyle and often leading to mortality. Keeping track of blood glucose levels noninvasively has been made possible due to diverse breakthroughs in wearable sensor technology coupled with holistic digital healthcare. Efficient glucose management has been revolutionized by the development of continuous glucose monitoring sensors and wearable, non/minimally invasive devices that measure glucose concentration by exploiting different physical principles, e.g., glucose oxidase, fluorescence, or skin dielectric properties, and provide real-time measurements every 1–5 min. This paper presents a highly novel and completely non-invasive sweat sensor platform technology that can measure and report glucose concentrations from passively expressed human eccrine sweat using electrochemical impedance spectroscopy and affinity capture probe functionalized sensor surfaces. The sensor samples 1–5 µL of sweat from the wearer every 1–5 min and reports sweat glucose from a machine learning algorithm that samples the analytical reference values from the electrochemical sweat sensor. These values are then converted to continuous time-varying signals using the interpolation methodology. Supervised machine learning, the decision tree regression algorithm, shows the goodness of fit R2 of 0.94 was achieved with an RMSE value of 0.1 mg/dL. The output of the model was tested on three human subject datasets. The results were able to capture the glucose progression trend correctly. Sweet sensor platform technology demonstrates a dynamic response over the physiological sweat glucose range of 1–4 mg/dL measured from 3 human subjects. The technology described in the manuscript shows promise for real-time biomarkers such as glucose reporting from passively expressed human eccrine sweat.

Abstract 9419: Quantification and Novel Algorithm of Echocardiography Against Magnetic Resonance Imaging for Isolated Tricuspid Regurgitation

Background: Transthoracic echocardiography (TTE) is the first-line imaging modality to evaluate isolated tricuspid regurgitation (TR), however comparisons with magnetic resonance imaging (MRI) measurements have seldomly been studied. We assessed the correlation, discriminative ability and derived a novel TTE-based algorithm using TR and right heart parameters to identify isolated severe TR by MRI. Methods: Isolated TR patients graded at least moderate-to-severe by TTE and undergoing MRI between January 2007-June 2019 were studied. TR and right heart parameters were analyzed by correlation, receiver-operative characteristics analysis and classification and regression tree algorithm for detection of severe TR by MRI (regurgitant volume &gt; 45mL and/or fraction &gt; 50%). Results: In 262 isolated TR patients, 108 (41.2%) had severe TR by MRI. There were moderate correlations and discrimination for most TTE-derived TR parameters and indexed right atrial volume (Pearson r=0.428-0.645) for TR quantification by MRI. This included a mean absolute and percentage difference of +10mL and +19% for regurgitant volume by TTE higher than MRI. However, there were none to modest correlations between TTE-derived right ventricle parameters with MRI-derived TR or right ventricle quantification. Key TTE parameters to identify severe TR by MRI in the decision tree regression algorithm were right atrial volume indexed &gt;47cm 2 /m 2 , then effective regurgitant orifice area &gt;0.45 cm 2 , and finally those with right ventricle free wall strain &gt;-9.5% had the highest risk. Conclusion: TTE-derived TR and right atrial quantification had moderate correlation and discrimination of severe TR by CMR, from which a novel TTE-based decision tree algorithm was derived, while RV correlation was relatively poor between TTE and CMR.

Study of Ventilator Breathing Parameters with Regards to Performance in Bag-Valve-Mask Compression Using ANN and Boosted Decision Tree Regression Algorithms

Study of Ventilator Breathing Parameters with Regards to Performance in Bag-Valve-Mask Compression Using ANN and Boosted Decision Tree Regression Algorithms

Artificial Intelligence Algorithms for the Analysis of Mechanical Property of Friction Stir Welded Joints by using Python Programming

In modern computational science, the interplay existing between machine learning and optimization process marks the most vital developments. Optimization plays an important role in mechanical industries because it leads to reduce in material cost, time consumption and increase in production rate. The recent work focuses on performing the optimization task on Friction Stir Welding process for obtaining the maximum Ultimate Tensile Strength (UTS) of the friction stir welded joints. Two machine learning algorithms i.e. Artificial Neural Network (ANN) and Decision Trees regression model are selected for the purpose. The input variables are Tool Rotational Speed (RPM), Tool Traverse Speed (mm/min) and Axial Force (KN) while the output variable is Ultimate Tensile Strength (MPa). It is observed that in case of the Artificial Neural Networks the Root Mean Square Errors for training and testing sets are 0.842 and 0.808 respectively while in case of Decision Trees regression model, the training and testing sets result Root Mean Square Errors of 11.72 and 14.61. So, it can be concluded that ANN algorithm gives better and accurate result than Decision Tree regression algorithm.