ترغب بنشر مسار تعليمي؟ اضغط هنا

A Bayesian Machine Learning Algorithm for Predicting ENSO Using Short Observational Time Series

241   0   0.0 ( 0 )
 نشر من قبل Faheem Gilani
 تاريخ النشر 2021
  مجال البحث فيزياء
والبحث باللغة English




اسأل ChatGPT حول البحث

A simple and efficient Bayesian machine learning (BML) training and forecasting algorithm, which exploits only a 20-year short observational time series and an approximate prior model, is developed to predict the Ni~no 3 sea surface temperature (SST) index. The BML forecast significantly outperforms model-based ensemble predictions and standard machine learning forecasts. Even with a simple feedforward neural network, the BML forecast is skillful for 9.5 months. Remarkably, the BML forecast overcomes the spring predictability barrier to a large extent: the forecast starting from spring remains skillful for nearly 10 months. The BML algorithm can also effectively utilize multiscale features: the BML forecast of SST using SST, thermocline, and wind burst improves on the BML forecast using just SST by at least 2 months. Finally, the BML algorithm also reduces the forecast uncertainty of neural networks and is robust to input perturbations.

قيم البحث

اقرأ أيضاً

Smooth power generation from solar stations demand accurate, reliable and efficient forecast of solar energy for optimal integration to cater market demand; however, the implicit instability of solar energy production may cause serious problems for t he smooth power generation. We report daily prediction of solar energy by exploiting the strength of machine learning techniques to capture and analyze complicated behavior of enormous features effectively. For this purpose, dataset comprising of 98 solar stations has been taken from energy competition of American Meteorological Society (AMS) for predicting daily solar energy. Forecast models of base line regressors including linear, ridge, lasso, decision tree, random forest and artificial neural networks have been implemented on the AMS solar dataset. Grid size is converted into two sections: 16x9 and 10x4 to ascertain attributes contributing more towards the generated power from densely located stations on global ensemble forecast system (GEFS). To evaluate the models, statistical measures of prediction error in terms of RMSE, MAE and R2_score have been analyzed and compared with the existing techniques. It has been observed that improved accuracy is achieved through random forest and ridge regressor for both grid sizes in contrast to all other proposed methods. Stability and reliability of the proposed schemes are evaluated on a single solar station as well as on multiple independent runs.
We develop a framework for analyzing multivariate time series using topological data analysis (TDA) methods. The proposed methodology involves converting the multivariate time series to point cloud data, calculating Wasserstein distances between the persistence diagrams and using the $k$-nearest neighbors algorithm ($k$-NN) for supervised machine learning. Two methods (symmetry-breaking and anchor points) are also introduced to enable TDA to better analyze data with heterogeneous features that are sensitive to translation, rotation, or choice of coordinates. We apply our methods to room occupancy detection based on 5 time-dependent variables (temperature, humidity, light, CO2 and humidity ratio). Experimental results show that topological methods are effective in predicting room occupancy during a time window. We also apply our methods to an Activity Recognition dataset and obtained good results.
Predicting the outcome of a chemical reaction using efficient computational models can be used to develop high-throughput screening techniques. This can significantly reduce the number of experiments needed to be performed in a huge search space, whi ch saves time, effort and expense. Recently, machine learning methods have been bolstering conventional structure-activity relationships used to advance understanding of chemical reactions. We have developed a model to predict the products of catalytic reactions on the surface of oxygen-covered and bare gold using machine learning. Using experimental data, we developed a machine learning model that maps reactants to products, using a chemical space representation. This involves predicting a chemical space value for the products, and then matching this value to a molecular structure chosen from a database. The database was developed by applying a set of possible reaction outcomes using known reaction mechanisms. Our machine learning approach complements chemical intuition in predicting the outcome of several types of chemical reactions. In some cases, machine learning makes correct predictions where chemical intuition fails. We achieve up to 93% prediction accuracy for a small data set of less than two hundred reactions.
We employ the framework of the Koopman operator and dynamic mode decomposition to devise a computationally cheap and easily implementable method to detect transient dynamics and regime changes in time series. We argue that typically transient dynamic s experiences the full state space dimension with subsequent fast relaxation towards the attractor. In equilibrium, on the other hand, the dynamics evolves on a slower time scale on a lower dimensional attractor. The reconstruction error of a dynamic mode decomposition is used to monitor the inability of the time series to resolve the fast relaxation towards the attractor as well as the effective dimension of the dynamics. We illustrate our method by detecting transient dynamics in the Kuramoto-Sivashinsky equation. We further apply our method to atmospheric reanalysis data; our diagnostics detects the transition from a predominantly negative North Atlantic Oscillation (NAO) to a predominantly positive NAO around 1970, as well as the recently found regime change in the Southern Hemisphere atmospheric circulation around 1970.
The Anderson Impurity Model (AIM) is a canonical model of quantum many-body physics. Here we investigate whether machine learning models, both neural networks (NN) and kernel ridge regression (KRR), can accurately predict the AIM spectral function in all of its regimes, from empty orbital, to mixed valence, to Kondo. To tackle this question, we construct two large spectral databases containing approximately 410k and 600k spectral functions of the single-channel impurity problem. We show that the NN models can accurately predict the AIM spectral function in all of its regimes, with point-wise mean absolute errors down to 0.003 in normalized units. We find that the trained NN models outperform models based on KRR and enjoy a speedup on the order of $10^5$ over traditional AIM solvers. The required size of the training set of our model can be significantly reduced using furthest point sampling in the AIM parameter space, which is important for generalizing our method to more complicated multi-channel impurity problems of relevance to predicting the properties of real materials.
التعليقات
جاري جلب التعليقات جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
mircosoft-partner

هل ترغب بارسال اشعارات عن اخر التحديثات في شمرا-اكاديميا