Auswahl der wissenschaftlichen Literatur zum Thema „Multioutput regression“

Cable force monitoring is an essential and critical part of structural health monitoring for cable-supported bridges. The quality of obtained information depends considerably on the number and location of limited sensors. The purpose of this article is to provide a method for optimal sensor placement for cable force monitoring in cable-supported bridges. Based on the spatial correlation between neighbouring or symmetrical cable forces, the structural information of non-monitored cables can be predicted by multioutput support vector regression models, established between monitored (input) and the non-monitored (output) cable forces. The number and placement of cable force sensors have significant influence on prediction performance of established multioutput support vector regression models. The proposed optimal sensor configuration is to select multioutput support vector regression models with minimum prediction error from all possible sensor locations. In this study, information entropy is employed to measure the prediction performance of different sensor configurations and formulate the objective function, optimised by three computationally effective algorithms: forward sequential sensor placement algorithm, backward sequential sensor placement algorithm and genetic algorithm. The application of proposed method to Nanjing No. 3 Yangtze River Bridge confirmed the efficiency, accuracy and effectiveness of the proposed method.

In a construction project, construction cost estimation is very important, but construction costs are affected by various factors, so they are difficult to predict accurately. However, with the recent development of ANN technology, it has become possible to predict construction costs with consideration of various influencing factors. Unlike previous research cases, this study aimed to predict the total construction cost by predicting seven sub-construction costs using a multioutput regression model, not by predicting a single total construction cost. In addition, analysis of the change in construction cost prediction performance was conducted by scaling and regularization. We estimated the error rate of predicting construction costs through sub-construction cost prediction to be 16.80%, a level similar to that of the total construction cost prediction error rate of 17.67%. This study shows that the construction cost can be calculated by predicting detailed cost factors at once, and it is expected that various types of construction costs or partial construction costs can be predicted using the predicted detailed cost elements. As a result of predicting several sub-construction costs using multioutput-based ANN, it was found that the prediction error rate varies depending on the type of construction. To improve accuracy, it is necessary to supplement influencing factors suitable for the construction features.

The displacement of concrete dams effectively reflects their structural integrity and operational status. Therefore, establishing a model for predicting the displacement of concrete dams and studying the evolution mechanism of dam displacement is essential for monitoring the structural safety of dams. Current data-driven models utilize artificial data that cannot reflect the actual status of dams for network training. They also have difficulty extracting the temporal patterns from long-term dependencies and obtaining the interactions between the targets and variables. To address such problems, we propose a novel model for predicting the displacement of dams based on the temporal convolutional network (TCN) with the attention mechanism and multioutput regression branches, named MLA-TCN (where MLA is multioutput model with attention mechanism). The attention mechanism implements information screening and weight distribution based on the importance of the input variables. The TCN extracts long-term temporal information using the dilated causal convolutional network and residual connection, and the multioutput regression branch achieves simultaneous multitarget prediction by establishing multiple regression tasks. Finally, the applicability of the proposed model is demonstrated using data on a concrete gravity dam within 14 years, and its accuracy is validated by comparing it with seven state-of-the-art benchmarks. The results show that the MLA-TCN model, with a mean absolute error (MAE) of 0.05 mm, a root-mean-square error (RMSE) of 0.07 mm, and a coefficient of determination (R2) of 0.99, has a comparably high predictive capability and outperforms the benchmarks, providing an accurate and effective method to estimate the displacement of dams.

Aiming at the problems of poor data effectiveness, low modeling accuracy, and weak generalization in the tuning process of microwave cavity filters, a parametric model for coaxial cavity filter using kernel canonical correlation analysis (KCCA) and multioutput least squares support vector regression (MLSSVR) is proposed in this study. First, the low-dimensional tuning data is mapped to the high-dimensional feature space by kernel canonical correlation analysis, and the nonlinear feature vectors are fused by the kernel function; second, the multioutput least squares support vector regression algorithm is used for parametric modeling to solve the problems of low accuracy and poor prediction performance; third, the support vector of the parameter model is optimized by the differential evolution whale algorithm (DWA) to improve the convergence and generalization ability of the model in actual tuning. Finally, the tuning experiments of two cavity filters with different topologies are carried out. The experimental results show that the proposed method has an obvious improvement in generalization performance and prediction accuracy compared with the traditional methods.

The interferometer is a widely used direction-finding system with high precision. When there are comprehensive disturbances in the direction-finding system, some scholars have proposed corresponding correction algorithms, but most of them require hypothesis based on the geometric position of the array. The method of using machine learning that has attracted much attention recently is data driven, which can be independent of these assumptions. We propose a direction-finding method for the interferometer by using multioutput least squares support vector regression (MLSSVR) model. The application of this method includes the following: the construction of MLSSVR model training data, training and construction of the MLSSVR model, and the estimation of direction of arrival. Finally, the method is verified through numerical simulation. When there are comprehensive deviations in the system, the direction-finding accuracy can be effectively improved.

De nombreux indicateurs de performance existent en sport et en santé (guérison, réhabilitation, etc.) qui permettent de caractériser différents critères sportifs et thérapeutiques.Ces différents types de performance dépendent généralement de la charge de travail (ou de rééducation) subie par les sportifs ou patients.Ces dernières années, beaucoup d'applications de l'apprentissage automatique au sport et à la santé ont été proposées.La prédiction, voir l'explication de performances à partir de données de charges pourrait permettre d'optimiser les entraînements ou les thérapies.Dans ce contexte la gestion des données manquantes et l'articulation entre les types de charges et les différents indicateurs de performance considérés représentent les 2 problématiques traitées dans ce manuscrit à travers 4 applications. Les 2 premières concernent la gestion des données manquantes par une modélisation incertaine réalisée sur (i) des données de football professionnel fortement incomplètes et (ii) des données COVID-19 bruitées artificiellement. Pour ces 2 contributions, nous avons associé des modèles d'incertitude crédibilistes, textit{i.e.} basés sur la théorie des fonctions de croyance, à différentes méthodes d'imputation adaptées au contexte chronologique des entraînements/matchs et des thérapies.Une fois les données manquantes imputées sous formes de fonctions de croyance, le modèle crédibiliste des $k$ plus proches voisins adapté à la régression a été utilisé de manière à tirer profit des modèles d'incertitudes incertains associés aux données manquantes. Dans un contexte de prédiction de performances en match de handball en fonction des charges de travail passées, des modèles de régression multisorties sont utilisés pour prédire simultanément 7 indicateurs de performance athlétiques et techniques. La dernière application concerne la rééducation de patients post-AVC ayant partiellement perdu l'usage d'un bras. De manière à détecter les patients non-répondant à la thérapie, le problème de la prédiction de différents critères de réhabilitation a permis de réappliquer les différentes contributions de ce manuscrit (imputation crédibiliste de données manquantes et régression multisorties pour la prédiction simultanée de différents indicateurs de performance
Numerous performance indicators exist in sport and health (recovery, rehabilitation, etc.), allowing us to characterize different sporting and therapeutic criteria.These different types of performance generally depend on the workload (or rehabilitation) undergone by athletes or patients.In recent years, many applications of machine learning to sport and health have been proposed.Predicting or even explaining performance based on workload data could help optimize training or therapy.In this context, the management of missing data and the articulation between load types and the various performance indicators considered represent the 2 issues addressed in this manuscript through 4 applications. The first 2 concern the management of missing data through uncertain modeling performed on (i) highly incomplete professional soccer data and (ii) artificially noisy COVID-19 data. For these 2 contributions, we have combined credibilistic uncertainty models, based on the theory of belief functions, with various imputation methods adapted to the chronological context of training/matches and therapies.Once the missing data had been imputed in the form of belief functions, the credibilistic $k$ nearest-neighbor model adapted to regression was used to take advantage of the uncertain uncertainty patterns associated with the missing data. In the context of predicting performance in handball matches as a function of past workloads, multi-output regression models are used to simultaneously predict 7 athletic and technical performance indicators. The final application concerns the rehabilitation of post-stroke patients who have partially lost the use of one arm. In order to detect patients not responding to therapy, the problem of predicting different rehabilitation criteria has enabled the various contributions of this manuscript (credibilistic imputation of missing data and multiscore regression for the simultaneous prediction of different performance indicators

Satisfied surface topography is important to achieve the function of a part, thereby machined surface prediction is essential. A surface forecasting model called space-time multioutput support vector regression (STMSVR) is developed in this paper. With machined surfaces pervading in manufacturing, high definition metrology (HDM) is adopted to measure the three dimensional machined surface. Millions of data points are generated to represent the entire surface. The STMSVR model captures the spatial-temporal characteristics of the successively machined surface and predicts the future surface. To verify the prediction accuracy of STMSVR, a case study on the engine cylinder block face milling process is applied. The results indicate that the developed model achieves a good agreement between the predicted surface and the real surface using four important indexes.

Abstract Quantitative determination of mineralogy can be done using high-definition spectroscopic logging methods, however these methods are rarely used due to complexity and cost. Also, it is difficult to obtain mineralogical composition in unconventional formations due to presence of kerogen and high heterogeneity and anisotropy of such formations. This problem can be resolved by utilizing Machine Learning algorithms based on well logging and thermal profiling data which can improve and speed up reservoir characterisation. Special wrappers such as Multioutput Regressor and Regressor Chain were applied to test several machine learning models and strategies of well logs combinations on multiscale data from an unconventional formation in West Siberia to predict mass and volumetric fractions of minerals obtained from Litho Scanner. RMSE and MAE were used as regression metrics. To validate the results, theoretical model was used to calculate thermal conductivity based on mineral volume fractions and compared with experimental data. Regressor Chain showed better performance for weight fractions prediction when data was scarce. The Gradient Boosting Regressor encapsulated within a Regressor Chain exhibited the most favorable outcomes in relation to the precision of mapping. The evaluation contrasting the ML-based model with the LithoScanner exhibited an average discrepancy of 0.026, as measured by the RMSE metric.