Jozef Kiseľák

Jozef Kiseľák

Auto-generated profile, based on publicly available information

Jozef Kiseľák

Institute of Mathematics, Faculty of Science, P.J.Šafárik University in Košice, Kosice, Slovak Republic

Research Focus & Metrics

Academic Career Length

19Years (2008-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
25 July 2020 in Neural Computing and Applications
Verified
Journal Article
Open access
PDF file
JKJozef Kiseľák
Jozef Kiseľák
YLYing Lu
Ying Lu
JSJan Svihra
Jan Svihra
PSPeter Szépe
Peter Szépe
MSMilan Stehlik
Milan Stehlik
We address the following problem: given a set of complex images or a large database, the numerical and computational complexity and quality of approximation for neural network may drastically differ from one activation function to another. A general novel methodology, scaled polynomial constant unit activation function “SPOCU,” is introduced and shown to work satisfactorily on a variety of problems. Moreover, we show that SPOCU can overcome already introduced activation functions with good properties, e.g., SELU and ReLU, on generic problems. In order to explain the good properties of SPOCU, we provide several theoretical and practical motivations, including tissue growth model and memristive cellular nonlinear networks. We also provide estimation strategy for SPOCU parameters and its relation to generation of random type of Sierpinski carpet, related to the [pppq] model. One of the attractive properties of SPOCU is its genuine normalization of the output of layers. We illustrate SPOCU methodology on cancer discrimination, including mammary and prostate cancer and data from Wisconsin Diagnostic Breast Cancer dataset. Moreover, we compared SPOCU with SELU and ReLU on large dataset MNIST, which justifies usefulness of SPOCU by its very good performance.

1 September 2008 in Statistics & Probability Letters
Verified
Journal Article
Closed access
JKJozef Kiseľák
Jozef Kiseľák
MSMilan Stehlik
Milan Stehlik

Silvia VilcekovaSilvia Vilcekova
Silvia Vilcekova
ZAZoran Apostoloski
Zoran Apostoloski
LMLudmila Meciarova
Ludmila Meciarova
Eva BurdováEva Burdová
Eva Burdová
JKJozef Kiseľák
Jozef Kiseľák
People who live in buildings are exposed to harmful effects of indoor air pollution for many years. Therefore, our research is aimed to investigate the indoor air quality in family houses. The measurements of indoor air temperature, relative humidity, total volatile organic compounds (TVOC), particulate matters (PM) and sound pressure level were carried out in 25 houses in several cities of the Republic of Macedonia. Mean values of indoor air temperature and relative humidity ranged from 18.9 °C to 25.6 °C and from 34.1% to 68.0%, respectively. With regard to TVOC, it can be stated that excessive occurrence was recorded. Mean values ranged from 50 μg/m3 to 2610 μg/m3. Recommended value (200 μg/m3) for human exposure to TVOC was exceeded in 32% of houses. Mean concentrations of PM2.5 (particular matter with diameter less than 2.5 μm) and PM10 (diameter less than 10 μm) are determined to be from 16.80 μg/m3 to 30.70 μg/m3 and from 38.30 μg/m3 to 74.60 μg/m3 individually. Mean values of sound pressure level ranged from 29.8 dB(A) to 50.6 dB(A). Dependence between characteristics of buildings (Year of construction, Year of renovation, Smoke and Heating system) and data from measurements (Temperature, Relative humidity, TVOC, PM2.5 and PM10) were analyzed using R software. Van der Waerden test shows dependence of Smoke on TVOC and PM2.5. Permutational multivariate analysis of variance shows the effect of interaction of Renovation and Smoke.

LMLudmila Meciarova
Ludmila Meciarova
Silvia VilcekovaSilvia Vilcekova
Silvia Vilcekova
Eva BurdováEva Burdová
Eva Burdová
JKJozef Kiselák
Jozef Kiselák
Thirty five Slovak households were selected for an investigation of indoor environmental quality. Measuring of indoor air physical and chemical factors and a questionnaire survey was performed during May 2017. The range of permissible operative temperature was not met in 11% of objects. Relative humidity met the legislative requirements in all monitored homes. Concentrations of total volatile organic compounds (TVOCs) were significantly higher in the apartments than in the family houses. The average TVOC levels in the apartments and family houses were 519.7 µg/m3 and 330.2 µg/m3, respectively. Statistical analysis confirmed the effect of indoor air temperature, relative humidity and particulate matter (PM0.5 and PM1) on the levels of TVOCs. Higher TVOC levels were observed also in homes where it is not a common practice to open windows during cleaning activities. Other factors that had a statistically significant effect on concentrations of volatile organic compounds were heating type, attached garage, location of the apartment within residential building (the floor), as well as number of occupants. Higher TVOC concentrations were observed in indoor than outdoor environment, while further analysis showed the significant impact of indoor emission sources on the level of these compounds in buildings. The questionnaire study showed a discrepancy between objective measurement and subjective assessment in the household environment, and pointed to insufficient public awareness about volatile organic compounds (VOCs).

19 January 2018 in Neural Computing and Applications
Verified
Journal Article
Open access
PDF file
LSLedys Salazar
Ledys Salazar
ONOrietta Nicolis
Orietta Nicolis
FRFabrizio Ruggeri
Fabrizio Ruggeri
JKJozef Kisel’ák
Jozef Kisel’ák
MSMilan Stehlik
Milan Stehlik
In recent years, air pollution has been a major concern for its implications on human health. Specifically, ozone (\(\mathrm{O}_{3}\)) pollution is causing common respiratory diseases. In this paper, we illustrate the process of modeling and prediction hourly \(\mathrm{O}_ {3}\) pollution measurements using wavelet transforms. We split the time series of \(\mathrm{O}_{3}\) in daily intervals and estimate scale and wavelet coefficients for each interval by the discrete wavelet transform (DWT) with Haar filter. Subsequently we apply cumulated autoregressive integrated moving average (ARIMA) to estimate the coefficients and forecast their evolution in future intervals. Then the inverse discrete wavelet transform is implemented for the reconstruction of the time series and the forecast in the near future. In order to assess the performance of the proposed methodology, we compare the predictions obtained by the DWT–ARIMA with those obtained by the ARIMA model. Several theoretical results are shown through a simulation study.

1 March 2016 in Ecological Complexity
Verified
Journal Article
Closed access
MSMilan Stehlik
Milan Stehlik
JDJ. Dušek
J. Dušek
JKJ. Kiseľák
J. Kiseľák