Introduction: The present study discusses the importance of having a predictive method to determine the prognosis of patients with diseases like Covid-19. This method can assist physicians in making treatment decisions that improve survival rates and avoid unnecessary treatments. This research also highlights the importance of calibration, which is often overlooked in model evaluation. Without proper calibration, incorrect decisions can be made in disease treatment and preventive care. Therefore, the current study compares two highly accurate machine learning algorithms, Gradient Boosting and Extreme Gradient Boosting, not only in terms of prediction accuracy but also in terms of model calibration and speed. Methods: This study involved analyzing data from Covid-19 patients who were admitted to two hospitals in Mashhad city, Razavi Khorasan province, over a span of 18 months. The k-fold cross-validation method was employed on the training dataset (K=5) to conduct the study. The accuracy and calibration of two methods (Gradient Boosting and Extreme Gradient Boosting) in predicting survival were compared using the Concordance Index and calibration. Results: The Concordance Index values obtained for Gradient Boosting and Extreme Gradient Boosting models were 0. 734 and 0. 736, in the imbalanced and In the balanced data, the Concordance Index values were 0. 893 for Gradient Boosting and 0. 894 for Extreme Gradient Boosting. The surv. calib_beta index, the Gradient Boosting model had an estimated value of 0. 59 in the imbalanced data and 0. 66 in the balanced data. The Extreme Gradient Boosting model had an estimated value of 0. 86 in the balanced data and 0. 853 in the imbalanced data. The Extreme Gradient Boosting model was faster in the learning process compared to the Gradient Boosting model. Conclusion: The Gradient Boosting and Extreme Gradient Boosting models exhibited similar prediction accuracy and discrimination power, but the Extreme Gradient Boosting model demonstrated relatively good calibration compare to Gradient Boosting model.

AbstractThe aim of this study was to study the performance of xgboost algorithm in genomic evaluation of complex traits as an alternative for Gradient Boosting algorithm (GBM). To this end, genotypic matrices containing genotypic information for, respectively, 5,000 (S1), 10,000 (S2) and 50,000 (S3) single nucleotide polymorphisms (SNP) for 1000 individuals was simulated. Beside xgboost and GBM, the GBLUP which is known as an efficient algorithm in terms of accuracy, computing time and memory requirement was also used to predict genomic breeding values. xgboost, GBM and GBLUP were run in R software using xgboost, gbm and synbreed packages. All the analyses were done using a machine equipped with a Core i7-6800K CPU which had 6 physical cores. In addition, 32 gigabyte of memory was installed on the machine. The Person's correlation between predicted and true breeding values (rp,t) and the mean squared error (MSE) of prediction were computed to compare predictive performance of different methods. While GBLUP was the most efficient user of memory, GBM required a considerably high amount of memory to run. By increasing size of data from S1 to S3, GBM went out from the competition mainly due to its high demand for memory. Parallel computing with xgboost reduced running time by %99 compared to GBM. The speedup ratios (the ratio of the GBM runtime to the time taken by the parallel computing by xgboost) were 444 and 554 for the S1 and S2 scenarios, respectively. In addition, parallelization efficiency (speed up ratio/number of cores) were, respectively, 74 and 92 for the S1 and S2 scenarios, indicating that by increasing the size of data, the efficiency of parallel computing increased. The xgboost was considerably faster than GBLUP in all the scenarios studied. Accuracy of genomic breeding values predicted by xgboost was similar to those predicted by GBM. While the accuracy of prediction in terms of rp,t was higher for GBLUP, the MSE of prediction was lower for xgboost, specially for larger datasets. Our results showed that xgboost could be an efficient alternative for GBM as it had the same accuracy of prediction, was extremely fast and needed significantly lower memory requirement to predict the genomic breeding values.

Applications of machine learning techniques in concrete properties' prediction have great interest to many researchers worldwide. Indeed, some of the most common machine learning methods are those based on adopting Boosting algorithms. A new approach, Histogram-Based Gradient Boosting, was recently introduced to the literature. It is a technique that buckets continuous feature values into discrete bins to speed up the computations and reduce memory usage. Previous studies have discussed its efficiency in various scientific disciplines to save computational time and memory. However, the algorithm's accuracy is still unclear, and its application in concrete properties estimation has not yet been considered. This paper is devoted to evaluating the capability of Histogram-Based Gradient Boosting in predicting concrete's compressive strength and comparing its accuracy to other Boosting methods. Generally, the results of the study have shown that the Histogram-Based Gradient Boosting approach is capable of achieving reliable prediction of concrete compressive strength. Additionally, it showed the effects of each model's parameters on the accuracy of the estimation.

In this study, the differential pulse voltammetry (DPV) method was used to simultaneously determine bismuth and copper concentrations. A 25 bismuth and copper mixtures at the designed ratio were measured using the DPV technique. However, the overlapping differential pulse voltammograms obtained made it difficult to quantitatively analyze the concentrations based on adaptive peak current selection. To address this issue, the voltammograms were preprocessed using derivatization and peak subtraction. The second derivative voltammogram was found to be highly correlated with the copper-bismuth concentration ratio, resulting in improved fit and prediction accuracy. To further improve the accuracy and precision of the training and prediction results, XGBoost and Gradient Boosting regression models were applied. The XGBoost and Gradient Boosting regression models showed high accuracy and precision with r-squared values of 0.877 and 0.993 for copper, and 0.879 and 0.993 for bismuth, respectively. The mean recoveries of copper were 99.84% and 98.07%, while bismuth recoveries were 93.17% and 90.85% for XGBoost and Gradient Boosting, respectively. Additionally, cross-validation using 10 splits produced a mean score of 45.565 and a mean absolute error of 13.051 for copper, and a mean score of 13.600 and a mean absolute error of 10.920 for bismuth. Overall, the results indicate that the proposed method is an accurate and precise way to simultaneously determine bismuth and copper concentrations.

With the development of remote sensing science, the use of hyperspectral images is becoming more widespread. Classification is one of the most popular topics in hyperspectral remote sensing. In the last two decades, a number of methods have been proposed to address the problem of hyperspectral data classification.In the present study, a structure based on learning capsule networks has been used to classify hyperspectral images, so that the network structure can have the most optimal generation of features by using a convolution layer and a capsule layer, and at the same time Avoid overfitting the on training data. The obtained results show the high quality of production features in the proposed structure. With the development of remote sensing science, the use of hyperspectral images is becoming more widespread. Classification is one of the most popular topics in hyperspectral remote sensing. In the last two decades, a number of methods have been proposed to address the problem of hyperspectral data classification.In the present study, a structure based on learning capsule networks has been used to classify hyperspectral images, so that the network structure can have the most optimal generation of features by using a convolution layer and a capsule layer, and at the same time Avoid overfitting the on training data. The obtained results show the high quality of production features in the proposed structure.In order to improve the classification accuracy, the feature extraction approach through the designed network and the classification by the Extreme Gradient Boosting was compared with the classification method by the global deep network. The proposed capsule approach consists of 3 basic layers: 1) Prime caps, which are capsules of size 8 and 32 with 9 × 9 filters and movement step 2, 2) Digitcaps with 10 16-dimensional capsules, and 3) fully connected layer. The results of examining two approaches for deep networking as well as combining capsule networks with XGBoost reinforcement tree algorithm were compared. Approaches such as SVM, RF-200, LSTM, GRU and GRU-Pretanh were considered to compare the proposed approach based on the configurations mentioned in their research.Up in addition to the study and quality measurement of production vector deep features by the proposed method in different classifiers, the ability of deep global networks in the application of classification should also be examined. The results of examining two approaches for deep network and also combining CapsNet with XGBoost show that by using the proposed combined method, images are classified with 99% accuracy on training data and 97.5% accuracy on test data.Up in addition to the study and quality measurement of production vector deep features by the proposed method in different classifiers, the ability of deep global networks in the application of classification should also be examined.The results of examining two approaches for deep network and also combining CapsNet with XGBoost show that by using the proposed combined method, images are classified with 99% accuracy on training data and 97.5% accuracy on test data.