The paper presents a method for creating homogeneous and heterogeneous ensembles using three measures of regressor diversity, namely chi-square statistic, correlation coefficient and disagreement. Regressors included in the ensemble were selected from a pool of 100, 200 and 300 models generated using algorithms included in the WEKA data mining software package, such as M5P, M5Rules, Random Tree, and REPTree. Using three different measures, the degree of diversity of each pair of regressors was determined, and then 40 models with the highest diversity value were selected to construct ensembles. The obtained regressor ensembles were compared in terms of accuracy of prediction with models created using standard machine learning techniques such as linear regression, k-nearest neighbors, bagging and random forest. For comparative experiments, a real-world data set of purchase and sale transactions of residential real estate taken from a cadastral system was used. The research showed the usefulness of the proposed method for building ensembles.

This paper addresses usability problems of data entry techniques and patterns in web systems and web sites. In order to carry out a comparative analysis of the solutions a dedicated web application was developed as an experimental tool for performing usability testing. The application was designed in the form of a test divided into scenarios and exercises on filling up selected models. Furthermore, the application recorded usability metrics, questionnaires, open and closed questions, as well as users’ personal data. 40 participants, who took part remotely in the experiment, were split into two distinct groups and completed two separate tests that differed in input data and tested solutions. Some measures referring to the effectiveness, efficiency, and user satisfaction were collected. The results were thoroughly analyzed and the recommendations for the usage of particular models were formulated.

Machine learning models require numerous training examples to provide reliable predictions of real estate prices. Expert algorithms could be applied wherever only several training samples are available. The accuracy of two expert algorithms based on the sales comparison approach was experimentally examined using real-world data derived from a cadastral system and registry of real estate transactions. The performance of the algorithms was compared with three data driven regression models for property valuation. Statistical analysis of the obtained results was conducted.

The random subspace and random forest ensemble methods using a genetic fuzzy rule-based system as a base learning algorithm were developed in Matlab environment. The methods were applied to the real-world regression problem of predicting the prices of residential premises based on historical data of sales/purchase transactions. The computationally intensive experiments were conducted aimed to compare the accuracy of ensembles generated by the proposed methods with bagging, repeated holdout, and repeated cross-validation models. The statistical analysis of results was made employing nonparametric Friedman and Wilcoxon statistical tests.