A comparative study of Gaussian process regression with other three machine learning approaches in the performance prediction of centrifugal pump

Accurate prediction of performance indices using impeller parameters is of great importance for the initial and optimal design of centrifugal pump. In this study, a kernel-based non-parametric machine learning method named with Gaussian process regression (GPR) was proposed, with the purpose of predicting the performance of centrifugal pump with less effort based on available impeller parameters. Nine impeller parameters were defined as model inputs, and the pump performance indices, that is, the head and efficiency, were determined as model outputs. The applicability of three widely used nonlinear kernel functions of GPR including squared exponential (SE), rational quadratic (RQ) and Matern5/2 was investigated, and it was found by comparing with the experimental data that the SE kernel function is more suitable to capture the relationship between impeller parameters and performance indices because of the highest R square and the lowest values of max absolute relative error (MARE), mean absolute proportional error (MAPE), and root mean square error (RMSE). In addition, the results predicted by GPR with SE kernel function were compared with the results given by other three machine learning models. The comparison shows that the GPR with SE kernel function is more accurate and robust than other models in centrifugal pump performance prediction, and its prediction errors and uncertainties are both acceptable in terms of engineering applications. The GPR method is less costly in the performance prediction of centrifugal pump with sufficient accuracy, which can be further used to effectively assist the design and manufacture of centrifugal pump and to speed up the optimization design process of impeller coupled with stochastic optimization methods.

Interpolation Parameters in Inverse Distance-Weighted Interpolation Algorithm on DEM Interpolation Error

Although DEM occupies an important basic position in spatial analysis, so far, the quality of DEM modeling has still not reached a satisfactory accuracy. This research mainly discusses the influence of interpolation parameters in the inverse distance-weighted interpolation algorithm on the DEM interpolation error. The interpolation parameters to be studied in this paper are the number of search points, the search direction, and the smoothness factor. In order to study the optimization of IDW parameters, the parameters that have uncertain effects on DEM interpolation are found through analysis, such as the number of search points and smoothing factor. This paper designs an experiment for the optimization of the interpolation parameters of the polyhedral function and finds the optimal interpolation parameters through experimental analysis. Of course, the “optimum” here is not the only one, but refers to different terrain areas, which makes the interpolation results relatively good. The selection of search points will be one of the research focuses of this article. After determining the interpolation algorithm, the kernel function is also one of the important factors that affect the accuracy of DEM. The value of the smoothing factor in the kernel function has always been the focus of DEM interpolation research. Different terrains, different interpolations, and functions will have different optimal smoothing factors. The search direction is to ensure that the sampling points are distributed in all directions when the sampling points are sparse and to improve the contribution rate of the sampling points to the interpolation points. The selection of search shape is to improve computing efficiency and has no effect on DEM accuracy; the search radius is mainly controlled by the number of search points, and there are two methods: adaptive search radius and variable length search radius. When the weight coefficient k = 1 , 2 , 3 , 4 , the number of sampling points involved in the interpolation calculation is different, and the error in the residual varies greatly, and both increase with the increase of the number of sampling points in the parameter interpolation calculation. This research will help improve the quality evaluation of DEM.

Research on target tracking algorithm based on mean shift with adaptive bandwidth

The traditional mean shift algorithm used fixed kernels or symmetric kernel function, which will cause the target tracking lost or failure. The target tracking algorithm based on mean shift with adaptive bandwidth was proposed. Firstly, the signed distance constraint function was introduced to produce the anisotropic kernel function based on signed distance kernel function. This anisotropic kernel function satisfies that the value of the region function outside the target is zero, which provides accurate tracking window for the target tracking. Secondly, calculate the mean shift window center of anisotropic kernel function template, the theory basis is the sum of vector weights from the sample point in the tracking window to the center point is zero. Thirdly, anisotropic kernel function templates adaptive update implementation by similarity threshold to limit the change of the template between two sequential pictures, so as to realize real-time precise tracking. Finally, the contrast experimental results show that our algorithm has good accuracy and high real time.

Twin support vector machine using kernel function for colorectal cancer detection

Nowadays, machine learning technology is needed in the medical field. therefore, this research is useful for solving problems in the medical field by using machine learning. Many cases of colorectal cancer are diagnosed late. When colorectal cancer is detected, the cancer is usually well developed. Machine learning is an approach that is part of artificial intelligence and can detect colorectal cancer early. This study discusses colorectal cancer detection using twin support vector machine (SVM) method and kernel function i.e. linear kernels, polynomial kernels, RBF kernels, and gaussian kernels. By comparing the accuracy and running time, then we will know which method is better in classifying the colorectal cancer dataset that we get from Al-Islam Hospital, Bandung, Indonesia. The results showed that polynomial kernels has better accuracy and running time. It can be seen with a maximum accuracy of twin SVM using polynomial kernels 86% and 0.502 seconds running time.

Arabic Documents Classification by a Radial Basis Hybridization

In this paper, we propose a hybrid system for contextual and semantic indexing of Arabic documents, bringing an improvement to classical models based on n-grams and the Okapi model. This new approach takes into account the concept of the semantic vicinity of terms. We proceed in fact by the calculation of similarity between words using an hybridization of NGRAMs-OKAPI statistical measures and a kernel function in order to identify relevant descriptors. Terminological resources such as graphs and semantic dictionaries are integrated into the system to improve the indexing and the classification processes.

Multi-class classification for large datasets with optimized SVM by non-linear kernel function

Most important part of Support Vector Machines(SVM) are the kernels. Although there are several widely used kernel functions, a carefully designed kernel will help improve the accuracy of SVM. The proposed work aims to develop a new kernel function for a multi-class support vector machine, perform experiments on various data sets, and compare them with other classification methods. Directly it is not possible multiclass classification with SVM. In this proposed work first designed a model for binary class then extended with the one-verses-all approach. Experimental results have proved the efficiency of the new kernel function. The proposed kernel reduces misclassification and time. Other classification methods observed better results for some data sets collected from the UCI repository.

Wise Information Technology of Med: Human Pose Recognition in Elderly Care

The growing problem of aging has led to a social concern on how to take care of the elderly living alone. Many traditional methods based on visual cameras have been used in elder monitoring. However, these methods are difficult to be applied in daily life, limited by high storage space with the camera, low-speed information processing, sensitivity to lighting, the blind area in vision, and the possibility of revealing privacy. Therefore, wise information technology of the Med System based on the micro-Doppler effect and Ultra Wide Band (UWB) radar for human pose recognition in the elderly living alone is proposed to effectively identify and classify the human poses in static and moving conditions. In recognition processing, an improved PCA-LSTM approach is proposed by combing with the Principal Component Analysis (PCA) and Long Short Term Memory (LSTM) to integrate the micro-Doppler features and time sequence of the human body to classify and recognize the human postures. Moreover, the classification accuracy with different kernel functions in the Support Vector Machine (SVM) is also studied. In the real experiment, there are two healthy men and one woman (22–26 years old) selected to imitate the movements of the elderly and slowly perform five postures (from sitting to standing, from standing to sitting, walking in place, falling and boxing). The experimental results show that the resolution of the entire system for the five actions reaches 99.1% in the case of using Gaussian kernel function, so the proposed method is effective and the Gaussian kernel function is suitable for human pose recognition.