|
|||||
|
|
Forecasting the discomfort levels within the greater Athens area, Greece using artificial neural networks and multiple criteria analysis |
|
| Authors: | P A Vouterakos K P Moustris A Bartzokas I C Ziomas P T Nastos A G Paliatsos |
| Institution: | 1. School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh, UK 2. Department of Mechanical Engineering, Technological Education Institute of Piraeus, 250 Thivon and P. Ralli Str., 122 44, Athens, Greece 3. Laboratory of Meteorology, Department of Physics, University of Ioannina, 451 10, Ioannina, Greece 4. Laboratory of Process Analysis and Design, Department of Chemical Engineering, National Technical University of Athens, Zografou Campus, 157-80, Athens, Greece 5. Laboratory of Climatology and Atmospheric Environment, Faculty of Geology and Geoenvironment, University of Athens, Panepistimiopolis, 157-84, Athens, Greece 6. General Department of Mathematics, Technological Education Institute of Piraeus, 250 Thivon and P. Ralli Str., 122-44, Athens, Greece |
| Abstract: | In this work, artificial neural networks (ANNs) were developed and applied in order to forecast the discomfort levels due to the combination of high temperature and air humidity, during the hot season of the year, in eight different regions within the Greater Athens area (GAA), Greece. For the selection of the best type and architecture of ANNs-forecasting models, the multiple criteria analysis (MCA) technique was applied. Three different types of ANNs were developed and tested with the MCA method. Concretely, the multilayer perceptron, the generalized feed forward networks (GFFN), and the time-lag recurrent networks were developed and tested. Results showed that the best ANNs type performance was achieved by using the GFFN model for the prediction of discomfort levels due to high temperature and air humidity within GAA. For the evaluation of the constructed ANNs, appropriate statistical indices were used. The analysis proved that the forecasting ability of the developed ANNs models is very satisfactory at a significant statistical level of p?<?0.01. |
| Keywords: | |
| 本文献已被 SpringerLink 等数据库收录！ | |