Aiming to hide the real money gains and to avoid taxes, fictive prices are sometimes recorded in the real estate transactions. This paper is concerned with artificial neural networks based screening of real estate transactions aiming to categorize them into "clear" and "fictitious" classes. The problem is treated as an outlier detection task. Both unsupervised and supervised approaches to outlier detection are studied here. The soft minimal hyper-sphere support vector machine (SVM) based novelty detector is employed to solve the task without the supervision. In the supervised case, the effectiveness of SVM, multilayer perceptron (MLP), and a committee based classification of the real estate transactions are studied. To give the user a deeper insight into the decisions provided by the models, the real estate transactions are not only categorized into "clear" and "fictitious" classes, but also mapped onto the self organizing map (SOM), where the regions of "clear", "doubtful" and "fictitious" transactions are identified. We demonstrate that the stability of the regions evolved in the SOM during training is rather high. The experimental investigations performed on two real data sets have shown that the categorization accuracy obtained from the supervised approaches is considerably higher than that obtained from the unsupervised one. The obtained accuracy is high enough for the technique to be used in practice.
The text describes the optimization task of renewable energy sources distributed to electrical microgrid of fictitious intelligent area that consists of intelligent buildings. Firstly, to solve this task a general optimization heuristic method of simulated annealing will be described. Testing was performed on the analytical functions but those will be only covered marginally. Of the tests on the approximation functions the method of simulated annealing would be the most suitable algorithm for the optimization task. Furthermore, two experiments were introduced. The first lies in the application of cluster analysis on daily diagrams of electricity consumption in intelligent buildings. Because the modeled year history of hourly electricity consumption is represented by multidimensional data this data forms the training set during the adaptive dynamics submitted to a competence model of neural network by days. After the network adaptation process the Kohonen's map during the adaptive dynamics will be drawn, from which required clusters can be read. In the second experiment a sorting design of the resources for typical days of a week is performed in the computer program UniCon.