Techniques Tanimoto correlated feature selection system and hybridization of clustering and boosting ensemble classification of remote sensed big data for weather forecasting
تکنیک های مربوط به سیستم انتخاب ویژگی Tanimoto و ترکیبی از خوشه بندی و افزایش طبقه بندی گروه از داده های بزرگ از راه دور برای پیش بینی آب و هوا-2020
Weather forecasting has been done using various techniques but still not efficient for handling the big remote sensed data since the data comprises the more features. Hence the techniques degrade the forecasting accuracy and take more prediction time. To enhance the prediction accuracy (PA) with minimal time, Tanimoto Correlation based Combinatorial MAP Expected Clustering and Linear Program Boosting Classification (TCCMECLPBC) Technique is proposed. At first, the data and features are gathered from big weather database. After that, relevant features are selected through finding the similarity between the features. Tanimoto Correlation Coefficient is used to find the similarity between the features for selecting the relevant features with higher feature selection accuracy. After selecting the relevant features, MAP expected clustering process is carried out to group the weather data for cluster formation. In this process, a number of cluster and cluster centroids are initialized. In this clustering process, it includes two steps namely expectation (E) and maximization (M) to discover maximum probability for grouping data into the cluster. After that, the clustering result is given to Linear Program boosting classifier to improve the prediction performance. In this classification, the weak classifier results are boosted to create strong classifier. The results evident that the TC-CMECLPBC technique enhance the PA with lesser time and false positive rate (FPR) than the conventional methods.
Keywords: Big data | Tanimoto correlation | MAP expected | Boosting classification | Expectation | Maximization | Similarity | Clustering | Cluster centroids | Strong classifier | Weak classifier
Deep shared representation learning for weather elements forecasting
یادگیری نمایندگی اشتراکی عمیق برای پیش بینی عناصر آب و هوا-2019
The accuracy and reliability of weather forecasting are of importance for many economic, business and management activities. This paper introduces novel data-driven predictive models based on deep convolutional neural networks (CNN) architecture for temperature and wind speed prediction in weather data. In particular, the proposed deep learning framework employs different upgrading versions of the convolutional neural networks i.e. 1d-, 2d- and 3d-CNN. The introduced models exploit the spatio-temporal multivariate weather data for learning shared representations using historical data and forecasting weather elements for a number of user defined weather stations simultaneously in an end-to-end fashion. The embedded feature learning component of the models as well as coupling the learned features of different input layers have shown to have a significant impact on the prediction task. The proposed models show promising results compared to the classical neural networks architecture used for modeling nonlinear systems. Two experimental setups have been considered based on a dataset collected from the Weather Underground website at six stations located in Netherlands and Belgium as well as a larger dataset with higher temporal resolution from the National Climatic Data Center (NCDC) at five stations located in Denmark. First, we focus on simultaneously predicting the temperature of two main stations of Amsterdam and Brussels for 1–10 days ahead. The second experiment concerns wind speed prediction at three weather stations located in Denmark for 6 and 12 h ahead. The obtained numerical results show that learning new shared representations of the weather data by means of convolutional operations improves the prediction performance.
Keywords: Deep learning | Weather forecasting | Convolutional neural networks | Dimensionality reduction | Representation learning
A dynamic neural network architecture with immunology inspired optimization for weather data forecasting
یک معماری شبکه عصبی پویا با ایمنولوژی بهینه سازی برای پیش بینی داده های آب و هوایی-2018
Recurrent neural networks are dynamical systems that provide for memory capabilities to recall past behaviour, which is necessary in the prediction of time series. In this paper, a novel neural network architecture inspired by the immune algorithm is presented and used in the forecasting of naturally occurring signals, including weather big data signals. Big Data Analysis is a major research frontier, which attracts extensive attention from academia, industry and government, particularly in the context of handling issues related to complex dynamics due to changing weather conditions. Recently, extensive deployment of IoT, sensors, and ambient intelligence systems led to an exponential growth of data in the climate domain. In this study, we concentrate on the analysis of big weather data by using the Dynamic Self Organized Neural Network Inspired by the Immune Algorithm. The learning strategy of the network focuses on the local properties of the signal using a self-organised hidden layer inspired by the immune algorithm, while the recurrent links of the network aim at recalling previously observed signal patterns. The proposed network exhibits improved performance when compared to the feedforward multilayer neural network and state-of-the-art recurrent networks, e.g., the Elman and the Jordan networks. Three non-linear and non-stationary weather signals are used in our experiments. Firstly, the signals are transformed into stationary, followed by 5-steps ahead prediction. Improvements in the prediction results are observed with respect to the mean value of the error (RMS) and the signal to noise ratio (SNR), however to the expense of additional computational complexity, due to presence of recurrent links.
Keywords: Recurrent Neural Networks ،Immune Systems Optimisation، Time Series Data analytics ، weather forecasting
Systematic Design of an Ideal Toolflow for Accelerating Big Data Applications on FPGA Platforms
طراحی سیستماتیک یک گردش ابزار ایده آل برای سرعت بخشیدن به برنامه های کاربردی داده های بزرگ در پلت فرم های FPGA-2018
The tremendous explosion of data has led to the “big data challenge” in the various domains of the current digital age including financial analytics, weather forecasting and bioinformatics. The processing requirements of the voluminous and complex data sets produced by the current data explosion are outpacing the computational capacity of traditional hardware platforms and thus necessitating adoption of high performance computing architectures such as clusters, cloud computing and customisable processing hardware such as field programmable gate arrays. In particular, FPGAs offer excellent flexibility, massive parallel computational capacity and good power efficiency which can meet the high processing demands of big data applications. However, despite their excellent processing merits, FPGAs are still suffering from low adoption by designers. Standard FPGA languages and tools are difficult and exclusive to users with digital hardware design expertise. Multiple high-level languages and design flows targeted at different application domains have been developed to meet the FPGA design challenge. However, there is a lack of a standardised specification that defines clearly how a highlevel FPGA design flow should be and what it should be capable of. This paper employs a system engineering approach to design and prototype an ideal high-level FPGA design Toolflow for the computational finance domain which utilises a simple standard software programming language to program the FPGA. The detailed specification of the ideal high-level FPGA Toolflow is presented and discussed. Preliminary results between a purely software design in comparison to a hardware design generated using the prototyped high-level FPGA Toolflow are presented.
Keywords: field programmable gate arrays; high-level design; toolflows; big data, system engineering
Black carbon pollution for a major road in Beijing_ Implications for policy interventions of the heavy-duty truck fleet
آلودگی کربن سیاه برای جاده های بزرگ در پکن. پیامدهای مداخلات سیاستی ناوگان کامیون سنگین-2017
Increasing attentions have been attained to transportation air quality in developing countries. Multi-disciplinary models are useful to understand the relationship between traffic dynamics and air quality, however, which generate higher demand for input data regarding traffic and emis sion. In Beijing, the environmental impact from non-local heavy-duty trucks (HDTs) has not been evaluated yet, which may be an important reason for higher black carbon concentration at night. This study selected one link of the North Fourth Ring Road, a major urban traffic corridor, to characterize BC pollution for typical weekdays by coupling traffic data investigation, vehicle emission calculation and roadside concentration modelling. We found local meteorological conditions would play a more governing role than changes of vehicle emissions in elevating roadside BC concentration, in particular for the night hours when dispersion conditions were significantly worsened. AERMOD-modelled results were in a good agreement with simulta neously observed BC concentration data, which could be responsible for 17% of the total roadside BC concentration during the summer period. We further identified that approximately 30–40% of the total HDT fleet on the North Fourth Ring Road were non-local HDTs, which could increase total BC emissions by ∼60% and roadside BC concentration by ∼2 μg m−3 at night compared to a local HDTs exclusive scenario. In addition to stricter emission controls for non-local HDTs, policy-makers may consider allowing HDTs to drive within the urban area at noon under a de licate scheme based on fine-grained weather forecasting.
Keywords: Black carbon | Roadside pollution | Vehicle emissions | Traffic | Transportation management