PM10 particles impose significant risks to human health and the well-being of individuals in general. However, due to the complexity of the inner-correlations between influencing environmental factors, the holistic approach to predictive analytics of PM10 concentration levels is a challenging task yet to be undertaken. We base this study on the rationale that a prediction model is suitable for making accurate estimations involving knowledge about the hidden interactions that govern them. In addition to the model’s precision, it is, therefore, beneficial to provide a model that is interpretable, as this can assist in the decision about how and which prevention actions to take. For this purpose, a Genetic Algorithm is proposed that carries out multiple regression analysis by searching for the optimal fictional definition of a prediction model. As such, the obtained model is human interpretable, where the preliminary analysis conducted within this study proved its compliance with the existing studies, while the model itself proved to be considerably more accurate than the present state-of-the-art.
COBISS.SI-ID: 21921558
This paper considers the use of interpolative coding for lossless chain code compression. The most popular chain codes are used, including Freeman chain code in eight (F8) and four directions (F4), Vertex Chain Code (VCC), and three-orthogonal chain code (3OT). The whole compression pipeline consists of the Burrows–Wheeler transform, Move-To-Front transform and the interpolative coding, which was improved by FELICS and new ?-coding. The approach was compared with the state-of-the-art chain code compression algorithms. For VCC, 3OT and F4, the obtained results are slightly better than the existing approaches. However, an important improvement was achieved with F8 chain code, where the presented approach is considerably better.
COBISS.SI-ID: 21154838
With the growing urbanization and environmental concerns over buildings' energy consumption and carbon footprint, the demand for energy-efficient building design is greater than ever. This paper addresses these concerns by presenting a novel method for estimating and optimising the thermal load (i.e. total energy load for heating and cooling) of a building within a real environment, provided by high-resolution LiDAR data, while considering long-term climatological parameters, estimated direct and anisotropic diffuse irradiance, shadowing from surroundings, and terrain topography. In the optimisation part of the method, the building's design is optimised regarding the estimated thermal load. The estimation was validated with the well-established EnergyPlus software. In experiments, a rectangular building's design was optimised on a flat and urban dataset. The effect of a building's design parameters on thermal load was inspected as well. On average, the proposed method improved a building's net heat gain by over 103 kWh/m2 and reduced its thermal load by 234.18 kWh/m2 when compared with the initial building design.
COBISS.SI-ID: 20948246
Community detection is a key to understanding the structure of complex networks. Many community detection approaches have been proposed based on the modularity optimization. Algorithms that optimize one initial solution often get into local optima, but algorithms that simultaneously optimize a population of solutions have high computational complexity. To solve these problems, genetic algorithms improved by a local learning procedure known as memetic algorithms can be applied. We propose a memetic algorithm for community detection in networks, that exploits node entropy for local learning. Node entropy is easy to use to speed up the convergence of an evolutionary algorithm and to increase the quality of partitions, while it uses only the node’s neighborhood and does not require any threshold value. Moreover, this algorithm is slightly modified in order to avoid modularity function which suffers a resolution limit and, therefore, it may fail to detect small communities. We propose and use an entropy function as an optimization function and as criteria in grouping crossover operator. Experiments on real-world and synthetic networks illustrate that the proposed method can find natural partitions effectively.
COBISS.SI-ID: 23825160
Outside the main mountain ranges and high North and South regions, individual isolated very small glaciers are the only glacier remnants and exceptional high-mountain active geomorphosites, which can be used to represent climate change consequences first hand to the local general public. The isolated, very small Triglav glacier in Slovenia was used to represent 3D glacier area changes for the period 1829–2016, together with long-term meteorological changes. Spatio-temporal changes of the glacier were derived mainly from old images and postcards with the help of interactive orientation (monoplotting), which enables the acquisition of a 3D glacier boundary from a single image by using a modern detailed digital elevation model. Very intuitive 3D visualisation was prepared, which shows the spatio-temporal changes of the glacier area, together with changes in average annual temperature and maximum annual snow depth. The last two are presented by colour palettes, where red colours represent stages when temperatures or maximum snow depths deviate from long-term averages in a negative way, meaning accelerating the glacier area reduction. Blue colours are used for stages when these parameters deviate from long-term averages in a positive way, meaning preserving the glacier area. From this 3D visualisation, one can easily recognise which meteorological parameter is the most important for the Triglav glacier preservation; this is the maximum annual snow depth. Such kind of 3D visualisation has a great potential for promotion of other active or evolving passive geomorphosites too.
COBISS.SI-ID: 34451203