Volume 9 (2018): Edition 2 (July 2018)

Background Determining the location, boundaries and areas of land properties accurately in the land cadastre is essential. The named data are provided using coordinates, acquired from field measurements. Since 2008, the Slovenian land cadastre claims positioning in the national realization of the ETRS89, so the GNSS use is practically indispensable. Objectives: Contrary to real-time, we can change parameters in GNSS post-processing. The aim of this paper is to simulate different measurement conditions for GNSS in order to determine how to acquire the best possible coordinates for further use in land area calculation. Methods/Approach: Simulations of obstacles near points followed the increasing of the cut-off angle. Furthermore, shortening the observation interval resulted in different occupation duration. The final condition evaluation for coordinate quality acquisition followed from fuzzy logic. Results: The results show that for short baselines, occupation duration is the most important factor in acquiring high quality coordinates and avoiding the multipath. Differences in coordinates from specific strategies can sometimes exceed the tolerance and evidently affect the area calculation. Conclusions: The findings confirm that only good measurement conditions lead to high quality coordinates and well-defined areas of land properties, which are the fundamental factor in relation to the issues of property valuation and assessing land taxes or rents.

Background: In this paper the well-known risk measurement method Conditional Value-at-Risk (CVaR) is applied to the Croatian stock market to estimate the risk for 8 sectors in Croatia. The method and an appropriate backtesting are applied to the sample of 29 stocks grouped into 8 sectors for the three different periods: the period of economic growth 2006-2007, the crisis period 2008-2009 and the post-crisis period 2013-2014, characterized by long-term economic stagnation in Croatia. Objectives: The objective of this paper is to estimate the risk of 8 sectors on the Croatian stock market in three different economic periods and to identify whether the sectors that are risky during the crisis period are the same sectors that are risky in the period of economic growth and economic stagnation. Methods/Approach: The Conditional Value-at-Risk method and an appropriate backtesting are applied. Results: Empirical findings indicate that sectors that are risky in the period of economic growth are not the same sectors that are risky during the period of economic crisis or stagnation. Conclusions: In all the three periods, the least risky sectors were Hotel-management, Tourism, Food, and Staples Retailing. The Construction sector in all the three periods was among the riskiest sectors

Background: Since high-frequency data have become available, an unbiased volatility estimator, i.e. realized variance (RV) can be computed. Commonly used models for RV forecasting suffer from strong persistence with a high sensitivity to the returns distribution assumption and they use only daily returns. Objectives: The main objective is measurement and forecasting of RV. Two approaches are compared: Heterogeneous AutoRegressive model (HAR-RV) and Feedforward Neural Networks (FNNs). Even though HAR-RV-type models describe RV stylized facts very well, they ignore its nonlinear behaviour. Therefore, FNN-HAR-type models are developed. Methods/Approach: Firstly, an optimal sampling frequency with application to the DAX index is chosen. Secondly, in and out of sample predictions within HAR models and FNNs are compared using RMSE, AIC, the Wald test and the DM test. Weights of FNN-HAR-type models are estimated using the BP algorithm. Results: The optimal sampling frequency of RV is 10 minutes. Within HAR-type models, HAR-RV-J has better, but not significant, forecasting performances, while FNN-HAR-J and FNNLHAR- J have significantly better predictive accuracy in comparison to the FNN-HAR model. Conclusions: Compared to the traditional ones, FNN-HAR-type models are better in capturing nonlinear behaviour of RV. FNN-HAR-type models have better accuracy compared to traditional HAR-type models, but only on the sample data, whereas their out-of-sample predictive accuracy is approximately equal.

Background: In practical use of machine learning models, users may add new features to an existing classification model, reflecting their (changed) empirical understanding of a field. New features potentially increase classification accuracy of the model or improve its interpretability. Objectives: We have introduced a guideline for determination of the sample size needed to reliably estimate the impact of a new feature. Methods/Approach: Our approach is based on the feature evaluation measure ReliefF and the bootstrap-based estimation of confidence intervals for feature ranks. Results: We test our approach using real world qualitative business-tobusiness sales forecasting data and two UCI data sets, one with missing values. The results show that new features with a high or a low rank can be detected using a relatively small number of instances, but features ranked near the border of useful features need larger samples to determine their impact. Conclusions: A combination of the feature evaluation measure ReliefF and the bootstrap-based estimation of confidence intervals can be used to reliably estimate the impact of a new feature in a given problem

Background: Hierarchical functional regions (FRs) can be calculated using data on interactions between basic spatial units (BSUs) and a hierarchical aggregation procedure. However, the results depend on the selected system of initial BSUs. In spatial sciences, this is known as the zonation effect, which is one of the effects of the Modifiable Areal Unit Problem (MAUP). Objectives: In this paper, we analyse the influence of the zonation effect on a system of hierarchical functional regions. Methods/Approach: We compared two systems of hierarchical functional regions of Slovenia modelled by the Intramax aggregation procedure using the inter-municipal labour commuting flows for the same year, but for two different initial sets of municipalities. Besides, we have introduced a new measure to compare systems of hierarchical FRs. Results: The results show that the zonation effect has an influence on hierarchical functional regions. The clustering comparison measure suggested here is a metric measure, which is appropriate for comparing hierarchical FRs. Conclusions: The zonation effect has influence on hierarchical FRs. The clustering comparison measure suggested in this paper is easy to interpret, but it should be adjusted for the number of clusterings

Background: During major maintenance projects on offshore installations, flotels are often used to accommodate the personnel. A gangway connects the flotel to the installation. If the offshore conditions are unfavorable, the responsible operatives need to decide whether to lift (disconnect) the gangway from the installation. If this is not done, there is a risk that an uncontrolled autolift (disconnection) occurs, causing harm to personnel and equipment. Objectives: We present a decision support model, developed using the DEXi tool for multi-criteria decision making, which produces advice on whether to disconnect/connect the gangway from/to the installation. Moreover, we report on our development method and experiences from the process, including the efforts invested. An evaluation of the resulting model is also offered, primarily based on feedback from a small group of offshore operatives and domain experts representing the end user target group. Methods/Approach: The decision support model was developed systematically in four steps: establish context, develop the model, tune the model, and collect feedback on the model. Results: The results indicate that the decision support model provides advice that corresponds with expert expectations, captures all aspects that are important for the assessment, is comprehensible to domain experts, and that the expected benefit justifies the effort for developing the model. Conclusions: We find the results promising, and believe that the approach can be fruitful in a wider range of risk-based decision support scenarios. Moreover, this paper can help other decision support developers decide whether a similar approach can suit them

Background: Internet of Things (IoT), earth observation and big scientific experiments are sources of extensive amounts of sensor big data today. We are faced with large amounts of data with low measurement costs. A standard approach in such cases is a stream mining approach, implying that we look at a particular measurement only once during the real-time processing. This requires the methods to be completely autonomous. In the past, very little attention was given to the most time-consuming part of the data mining process, i.e. data pre-processing. Objectives: In this paper we propose an algorithm for data cleaning, which can be applied to real-world streaming big data. Methods/Approach: We use the short-term prediction method based on the Kalman filter to detect admissible intervals for future measurements. The model can be adapted to the concept drift and is useful for detecting random additive outliers in a sensor data stream. Results: For datasets with low noise, our method has proven to perform better than the method currently commonly used in batch processing scenarios. Our results on higher noise datasets are comparable. Conclusions: We have demonstrated a successful application of the proposed method in real-world scenarios including the groundwater level, server load and smart-grid data

Background: Bike-sharing programmes have become popular in a large number of cities in order to facilitate bicycle use. Determining the location of bike sharing stations is vital to success of these programmes. Objectives: In this paper, a case study is applied to the Gaziantep University campus in order to find possible locations of the stations for users (students). The purpose is to minimize the total walking distance. Methods/Approach: Set and maximal covering mathematical models are considered to decide on coverage capability of determined 20 demand points and 20 potential bike stations. Then, the mathematical models of P-center and P-median are used to build possible stations and to allocate demand points to the opened stations. Finally, an undesirable facility location model is used to find the bike stations, which have the maximum distance from demand nodes, and to eliminate them. Results: In computational results, it is clearly seen that the proposed approaches set the potential bike station covering all demand points. They also provide different solutions for the campus planners. Conclusions: The methodology outlined in this study can provide university administrators with a useful insight into locations of stations, and in this way, it contributes significantly to future planning of bike-sharing systems.

Background: Hidden economy presents a major concern for all national economies, particularly for those of developing countries. Objectives: In this work, methods for determination of the size of hidden economy are discussed. Particular attention is devoted to the methods using electricity consumption as an indicator (the Lackó method and the Kaufmann and Kaliberda method). Methods/Approach: The modified Lackó method adapted for a single country and the sophisticated Kaufmann and Kaliberda method have been used. Results: It has been shown that such methods are effective in measurement of the hidden economy extent in small open economies exposed to severe external influences. The article presents results for Macedonia and their comparison with results for Croatia, as a good role-model for other states in Western Balkans. Conclusions: Model methods involving energy consumption are particularly efficient in determination of the size of the hidden economic sector in small open economies as those of the Western Balkan countries

Background: Determining the location, boundaries and areas of land properties accurately in the land cadastre is essential. The named data are provided using coordinates, acquired from field measurements. Since 2008, the Slovenian land cadastre claims positioning in the national realization of the ETRS89, so the GNSS use is practically indispensable. Objectives: Contrary to real-time, we can change parameters in GNSS post-processing. The aim of this paper is to simulate different measurement conditions for GNSS in order to determine how to acquire the best possible coordinates for further use in land area calculation. Methods/Approach: Simulations of obstacles near points followed the increasing of the cut-off angle. Furthermore, shortening the observation interval resulted in different occupation duration. The final condition evaluation for coordinate quality acquisition followed from fuzzy logic. Results: The results show that for short baselines, occupation duration is the most important factor in acquiring high quality coordinates and avoiding the multipath. Differences in coordinates from specific strategies can sometimes exceed the tolerance and evidently affect the area calculation. Conclusions: The findings confirm that only good measurement conditions lead to high quality coordinates and well-defined areas of land properties, which are the fundamental factor in relation to the issues of property valuation and assessing land taxes or rents