مقاله انگلیسی تحلیل کلان داده ها در پیش بینی وضعیت آب و هوا
Abstract
Introduction
Background
Related Works and Motivation
Research Methodology
Classifcation of the Selected Approaches
Discussion
Open Issues and Future Trends
Conclusion and Limitation
References
Abstract
Weather forecasting, as an important and indispensable procedure in people’s daily lives, evaluates the alteration happening in the current condition of the atmosphere. Big data analytics is the process of analyzing big data to extract the concealed patterns and applicable information that can yield better results. Nowadays, several parts of society are interested in big data, and the meteorological institute is not excluded. Therefore, big data analytics will give better results in weather forecasting and will help forecasters to forecast weather more accurately. In order to achieve this goal and to recommend favorable solutions, several big data techniques and technologies have been suggested to manage and analyze the huge volume of weather data from diferent resources. By employing big data analytics in weather forecasting, the challenges related to traditional data management techniques and technology can be solved. This paper tenders a systematic literature review method for big data analytic approaches in weather forecasting (published between 2014 and August 2020). A feasible taxonomy of the current reviewed papers is proposed as technique-based, technology-based, and hybrid approaches. Moreover, this paper presents a comparison of the aforementioned categories regarding accuracy, scalability, execution time, and other Quality of Service factors. The types of algorithms, measurement environments, modeling tools, and the advantages and disadvantages per paper are extracted. In addition, open issues and future trends are debated.
introduction
Originally weather forecasting started in the nineteenth century [1, 2]. The analysis of atmospheric data, including temperature, radiation, air pressure, wind speed, wind direction, humidity, and rainfall, is defned as weather forecasting. In order to predict the weather, a high volume of data must be collected or generated. Furthermore, these data are disorganized. Thus, utilizing the weather data for predicting the weather is a complex task, and it contains too many changeable parameters. These parameters vary according to the weather conditions that change very fast. To propose an algorithm for weather forecasting, we should consider its particular characteristics, such as continuity, data intensity, and multidimensional and chaotic behaviors [3, 4]. Originally weather forecasting has been developed from a human-intensive task [5] to a computational process [6], and to this end, it requires high-tech equipment. There are various factors that can afect the precision of forecasts. Season, geographical location, the accuracy of input data, classifcations of weather, lead time, and validity time are some of these efective factors [7, 8].