Search results

This paper aims to explore detection of heating load of building by machine learning. Detection of heating load of building is very important in design of buildings due to efficient energy consumption.

In this study, detection of heating load of building based on effects of dry-bulb temperature, dew-point temperature, radiation, diffuse radiation and wind speed was analyzed. Machine learning approach was implemented for such a purpose.

The obtained results could be useful for future planning of heating load of buildings. Because the heating load of building is a very nonlinear phenomenon, it is suitable to use machine learning approach to avoid the nonlinearity of the system.

The obtained results could be used effectively in detection of heating load of buildings.

This paper aims to show the implementation in the terrestrial trunked radio (TETRA)-based sensor network. The publicly available data show that, in Serbia, the annual damage caused by hailstorms in the past seven years has been estimated almost at an average level of 40m of euros. As the amount of hail was not changed, the hail suppression system of the Republic of Serbia has to be improved, both technically and organizationally, to get better efficiency and protection and to reduce the damage.

In this paper, the authors show the implementation and improvements in the modern terrestrial trunked radio (TETRA)-based sensor network, and they propose the scientific use of sensors for remote control of automatic hail suppression rocket stations.

The authors’ idea is that TETRA should be used as an operational and official telecommunicating system for hail suppression activities units. A number of sensors, connected in a network, are used to maintain a high-quality functioning of this digital radio system, managed remotely and controlled either by operators or automatically.

The presented study with a real example attempts to explain as to how the system functions and how it can improve hail suppression activities.

This paper aims to deal with the previously unknown prediction accuracy of students’ activity pattern in a blended learning environment.

To extract the most relevant activity feature subset, different feature-selection methods were applied. For different cardinality subsets, classification models were used in the comparison.

Experimental evaluation oppose the hypothesis that feature vector dimensionality reduction leads to prediction accuracy increasing.

Improving prediction accuracy in a described learning environment was based on applying synthetic minority oversampling technique, which had affected results on correlation-based feature-selection method.

The major contribution of the research is the proposed methodology for selecting the optimal low-cardinal subset of students’ activities and significant prediction accuracy improvement in a blended learning environment.