Search

Regression algorithms are regularly used for improving the accuracy of satellite precipitation products. In this context, satellite precipitation and topography data are the predictor variables, and gauged-measured precipitation data are the dependent variables. Alongside this, it is increasingly recognised in many fields that combinations of algorithms through ensemble learning can lead to substantial predictive performance improvements. Still, a sufficient number of ensemble learners for improving the accuracy of satellite precipitation products and their large-scale comparison are currently missing from the literature. In this study, we work towards filling in this specific gap by proposing 11 new ensemble learners in the field and by extensively comparing them. We apply the ensemble learners to monthly data from the PERSIANN (Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks) and IMERG (Integrated Multi-satellitE Retrievals for GPM) gridded datasets that span over a 15-year period and over the entire contiguous United States (CONUS). We also use gauge-measured precipitation data from the Global Historical Climatology Network monthly database, version 2 (GHCNm). The ensemble learners combine the predictions of six machine learning regression algorithms (base learners), namely the multivariate adaptive regression splines (MARS), multivariate adaptive polynomial splines (poly-MARS), random forests (RF), gradient boosting machines (GBM), extreme gradient boosting (XGBoost) and Bayesian regularized neural networks (BRNN), and each of them is based on a different combiner. The combiners include the equal-weight combiner, the median combiner, two best learners and seven variants of a sophisticated stacking method. The latter stacks a regression algorithm on top of the base learners to combine their independent predictions. Its seven variants are defined by seven different regression algorithms, specifically the linear regression (LR) algorithm and the six algorithms also used as base learners. The results suggest that sophisticated stacking performs significantly better than the base learners, especially when applied using the LR algorithm. It also beats the simpler combination methods. [ABSTRACT FROM AUTHOR]

Gridded satellite precipitation datasets are useful in hydrological applications as they cover large regions with high density. However, they are not accurate in the sense that they do not agree with ground-based measurements. An established means for improving their accuracy is to correct them by adopting machine learning algorithms. This correction takes the form of a regression problem, in which the ground-based measurements have the role of the dependent variable and the satellite data are the predictor variables, together with topography factors (e.g., elevation). Most studies of this kind involve a limited number of machine learning algorithms and are conducted for a small region and for a limited time period. Thus, the results obtained through them are of local importance and do not provide more general guidance and best practices. To provide results that are generalizable and to contribute to the delivery of best practices, we here compare eight state-of-the-art machine learning algorithms in correcting satellite precipitation data for the entire contiguous United States and for a 15-year period. We use monthly data from the PERSIANN (Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks) gridded dataset, together with monthly earth-observed precipitation data from the Global Historical Climatology Network monthly database, version 2 (GHCNm). The results suggest that extreme gradient boosting (XGBoost) and random forests are the most accurate in terms of the squared error scoring function. The remaining algorithms can be ordered as follows, from the best to the worst: Bayesian regularized feed-forward neural networks, multivariate adaptive polynomial splines (poly-MARS), gradient boosting machines (gbm), multivariate adaptive regression splines (MARS), feed-forward neural networks and linear regression. [ABSTRACT FROM AUTHOR]

Over the last years deep learning methods have been shown to outperform previous state-of-the-art machine learning techniques in several fields, with computer vision being one of the most prominent cases. This review paper provides a brief overview of some of the most significant deep learning schemes used in computer vision problems, that is, Convolutional Neural Networks, Deep Boltzmann Machines and Deep Belief Networks, and Stacked Denoising Autoencoders. A brief account of their history, structure, advantages, and limitations is given, followed by a description of their applications in various computer vision tasks, such as object detection, face recognition, action and activity recognition, and human pose estimation. Finally, a brief overview is given of future directions in designing deep learning schemes for computer vision problems and the challenges involved therein. [ABSTRACT FROM AUTHOR]

Proposes an unsupervised video object segmentation and tracking algorithm based on an adaptable neural-network architecture. Neural-network retraining strategy; Unsupervised retraining set estimation in videoconference sequences; Specifications of the decision mechanism; Experimental results; Conclusion.

Presents information on a study which described a neural network architecture for recursive nonlinear traffic prediction and modeling of MPEG video sources. Investigation of the problem of online traffic modeling; Implementation of the neural network architecture; Results and discussion; Conclusions.

Presents information on a study which described an approach for improving the performance of neural-network classifiers in image recognition, segmentation or coding applications based on a retraining algorithm. Formulation of the problem; Optimal selection of the network retraining data; Decision mechanism for network retraining; Conclusions.

Deep learning techniques are used for capturing intricate structures of large-scale data by employing computational models of multiple processing layers that can learn and represent data with multiple levels of abstraction. Such methods can include Convolutional Neural Networks, stacked auto-encoders and Long-Short Term Memory (LSTM) architectures. LSTM networks are suitable for dealing with time-dependent data through mapping input to output sequences as, for instance, in language modeling and speech recognition. One application that has recently attracted considerable attention within the geodetic community is the possibility of applying these techniques to account for the adverse effects of the ionospheric delays on the GNSS satellite signals. Ionospheric delay depend on three main factors: (i) the total electron content (TEC), (ii) the frequency of the GNSS signals, and (iii) the angle at which the signal enters the ionospheric layer. The combination of multi-frequency GNSS measurements allows us to remove most of the ionospheric effects. However, ionospheric prediction models have to be applied for single-frequency ionospheric estimation in order to remove this effect. This can be done using Global Ionospheric Maps (GIMs) available from the International GNSS Service (IGS) and evaluating the TEC effect via a mapping function involving the azimuth and elevation angles of the GNSS signals. This paper deals with a modeling approach suitable for predicting the ionospheric delay at different locations of the IGS network stations using the LSTM networks. We also incorporate a Bayesian optimization method for selecting the best configuration parameters of the LSTM network, thus improving network's performance. LSTM architecture models long-range dependencies in time series, making it appropriate for ionospheric modeling in GNSS positioning. As experimental data we used actual GNSS observations from the global IGS network stations participating in the MGEX project that provides various satellite signals from the currently available multiple navigation satellite systems. Slant TEC data (STEC) were obtained using the available GPS/GLONASS/Galileo signals after processing with various techniques, such as Precise Point Positioning (PPP). The combination of multi-GNSS signals in PPP, allows us to investigate additional biases such as Differential Code Biases (DCBs) that are closely related to the ionospheric delays. Consequently, a basic LSTM network structure is created, having as minimum inputs the following: time epoch, GNSS signal azimuth and elevation angle, ionospheric delays for the previous and present observation epochs. The sequential GNSS observations are fed one by one in a chain-like structure, in a way that a single state is stored in a unique node with a single activation function in the network at one timestep and then is propagated to the next timestep. In this way, the LSTM algorithm is able to predict future ionospheric delay values. Topics to be discussed in the paper include the design of the LSTM network structure, the Bayesian optimization strategy, the training methods exploiting steepest descent algorithms, data analysis, as well as preliminary testing results of the ionospheric delay predictions as compared, for instance, with the IGS ionosphere products and the widely used Klobuchar model. [ABSTRACT FROM AUTHOR]