Krzysztof Gajowniczek | Szkoła Główna Gospodarstwa Wiejskiego (original) (raw)

Papers by Krzysztof Gajowniczek

In this study we compared incomes distributions in the USA for two subgroups (defined according t... more In this study we compared incomes distributions in the USA for two subgroups (defined according to sex or race). We utilized the quantile decomposition method to describe differences between the two distributions as a function of their quantiles. The analyzed objects are characterized by the set of attributes (education, age, etc.). We evaluate strength of the influence of the attributes onto the various parts of the incomes distributions. In such a way we evaluate income inequalities and their causes in two subgroups of people.

The aim of this article is to establish whether econophysics can cause a scientic revolution and ... more The aim of this article is to establish whether econophysics can cause a scientic revolution and fundamentally change the image of mainstream economics. Science development processes were carefully analysed by Kuhn, who even created a specic vocabulary for it. The most important phrases include paradigm and scientic community. When comparing the disciplinary matrices of econophysics and economics, it has to be stressed that despite the absolute compatibility of the goals of both sciences, econophysics is not as postulated from time to time a new econometric approach that entails the application of physics in studies of economics, but rather it is a scientic eld totally dierent from economics. The disproportion between the disciplinary matrices of both sciences regards such elements as symbolic generalisations, models, values, and exemplars. Therefore, it seems that progressive accumulation of knowledge in economics will reveal new anomalies as well as deepen existing ones, making a paradigm shift inevitable. A scientic revolution should be expected at an international level, and in such countries as Poland it will be external and forced. The reasons for that lie in psychology and history. In 1989, in Poland and in other post-socialist countries, a rapid change in the disciplinary matrix of economics occurred and involved the replacement of the socialist economic paradigm with the capitalist economic paradigm. Another scientic revolution of such nature is right around the corner and entails replacing the disciplinary matrix of economics with the transdisciplinary matrix of econophysics. Since Polish economists have tried very hard to resist such a great number of changes, the paradigm shift will require deep involvement and much work from young scholars.

This paper examines the data-mining and supervised based machine learning models for predicting 1... more This paper examines the data-mining and supervised based machine learning models for predicting 1-month ahead cooling load demand of an office building, including the primitive intention of enhancing the forecasting performance and the accuracy. The data-mining and supervised based machine learning models include; regression support vector machine, Gaussian process regression, scaled conjugate gradient, tree bagger, boosted tree, bagged tree, neural network, multiple linear regression and bayesian regularization. The external climate data, hours/day in a week, previous week load, previous day load and previous 24-h average load are applied as input parameters for these models. Whereas, the output of the models is the electrical power required for water source heat pump. A water source heat pump located in Beijing, China, is selected for examining 1-month ahead cooling load forecasting, i.e., from July 8 to August 7, 2016. In this paper, simulations are classified into three sessions: 7-days, 14-days and 1-month. The forecast performance is assessed by computing four performance indices such as mean square error, mean absolute error, root mean square error and mean absolute percentage error. The mean absolute percentage error for 7-days ahead cooling load prediction of the water source heat pump from data-mining based models, Gaussian process regression, tree bagger, boosted tree, bagged tree and multiple linear regression were 0.405%, 3.544%, 1.928%, 1.703% and 13.053% respectively. While, mean absolute percentage error of 7-days ahead forecasting in case of machine learning based models such as a regression support vector machine, Bayesian regularization, scaled conjugate gradient and neural network were 12.761%, 2.314%, 6.314%, 2.592% respectively. The percentage forecasting error index proved that the results of data-mining based models are more precise and similar to the existing machine learning models. The results also demonstrate that the better performance and efficiency in foreseeing the abnormal behaviour in forecasting and future cooling load demand in the building environment.

The paper presents the improved method for 24 hour ahead load forecasting applied to the individu... more The paper presents the improved method for 24 hour ahead load forecasting applied to the individual household data from smart metering system. In this approach we decompose a set of individual forecasts into basis latent components with destructive or constructive impact on the prediction. The main research problem in such model aggregation is proper identification of destructive components which can be treated as some noise factors. To assess the randomness of signals and thus their similarity to the noise we used a new variability measure which helps to compare the decomposed signals with some typical noise models. The experiments performed on individual household electricity consumption data with blind separation algorithms contributed to forecasts improvements.

—The Grade Correspondence Analysis (GCA) with posterior clustering and visualization is introduce... more —The Grade Correspondence Analysis (GCA) with posterior clustering and visualization is introduced and applied to smart meter data on individual household level. The main task of this analysis is to reveal the latent structure of electricity usage patterns and to propose a two dimensional segmentation taking into account the usage of selected home appliances and time of their usage. This provides the solutions applicable in smart metering systems that can support usage forecasting and contribute to higher energy awareness.

Smart metering is a quite new topic that has grown in importance all over the world and it appear... more Smart metering is a quite new topic that has grown in importance all over the world and it appears to be a remedy for rising prices of electricity. Forecasting electricity usage is an important task to provide intelligence to the smart gir d. Accurate forecasting will enable a utility provider to plan the resources and also to take control actions to balance the electricity supply and demand. The customers will benefit from metering solutions through greater understanding of their own energy consumption and future projections, allowing them to better manage costs of their usage. In this proof of concept paper, our contribution is twofold: (1) we deal with short term electricity load forecasting for 24 hours ahead, not on the aggregate but on the individual household level what fits into the stream of Residential Power Load Forecasting (RPLF) methods; (2) we utilized a set of household behavioral data which significantly improved the forecasts accuracy.

—Two algorithms for building classification trees, based on Tsallis and Rényi entropy, are propos... more —Two algorithms for building classification trees, based on Tsallis and Rényi entropy, are proposed and applied to customer churn problem. The dataset for modeling represents highly unbalanced proportion of two classes, which is often found in real world applications, and may cause negative effects on classification performance of the algorithms. The quality measures for obtained trees are compared for different values of α parameter.

In this work we analyze empirically customer churn problem from a physical point of view to provi... more In this work we analyze empirically customer churn problem from a physical point of view to provide objective, data driven and significant answers to support decision making process in business application. In particular, we explore different entropy measures applied to decision trees and assess their performance from the business perspective using set of model quality measures often used in business practice. Additionally, the decision trees are compared with logistic regression and two machine learning methods — neural networks and support vector machines.

Leveraging smart metering solutions to support energy efficiency on the individual household leve... more Leveraging smart metering solutions to support energy efficiency on the individual household level poses novel research challenges in monitoring usage and providing accurate load forecasting. Forecasting electricity usage is an especially important component that can provide intelligence to smart meters. In this paper, we propose an enhanced approach for load forecasting at the household level. The impacts of residents' daily activities and appliance usages on the power consumption of the entire household are incorporated to improve the accuracy of the forecasting model. The contributions of this paper are threefold: (1) we addressed short-term electricity load forecasting for 24 hours ahead, not on the aggregate but on the individual household level, which fits into the Residential Power Load Forecasting (RPLF) methods; (2) for the forecasting, we utilized a household specific dataset of behaviors that influence power consumption, which was derived using segmentation and sequence mining algorithms; and (3) an extensive load forecasting study using different forecasting algorithms enhanced by the household activity patterns was undertaken.

—Demand peaks in electrical power system cause serious challenges for energy providers as these e... more —Demand peaks in electrical power system cause serious challenges for energy providers as these events are typically difficult to foresee and require the grid to support extraordinary consumption levels. Accurate peak forecasting enables a utility provider to plan the resources and also to take control actions to balance the supply and the demand of electricity. However, this is difficult in practice as it requires precision in prediction of peaks in advance. In this paper, our contribution is the proposal of data mining scheme to detect the peak load in the electricity system at country level. For this purpose we undertake the approach different from time series forecasting and represent it as a classical pattern recognition problem. We utilize set of machine learning techniques to benefit from accurate detection of the peaks in the Polish electric power system. The key finding is that the algorithms can accurately detect 96.2% of the electricity peaks up to 24 hours ahead

—The Grade Correspondence Analysis (GCA) with posterior clustering and visualization is introduce... more —The Grade Correspondence Analysis (GCA) with posterior clustering and visualization is introduced and applied to individual households' electricity usage data. The main task of this analysis is to identify a way of representing the variability of a households behavior and to develop an efficient way of clustering the households into a few, usable and homogenous groups. The regularity in terms of the electricity usage is useful information for organizations to allow accurate demand planning with the aim of improving the overall efficiency of the network. The approach is tested using data from 46 households located in Austin, Texas, USA and monitored for 14 months at a sampling interval of 1 hour.

Forecasting of electricity demand has become one of the most important areas of research in the e... more Forecasting of electricity demand has become one of the most important areas of research in the electric power industry, as it is a critical component of cost-efficient power system management and planning. In this context, accurate and robust load forecasting is supposed to play a key role in reducing generation costs, and deals with the reliability of the power system. However, due to demand peaks in the power system, forecasts are inaccurate and prone to high numbers of errors. In this paper, our contributions comprise a proposed data-mining scheme for demand modeling through peak detection, as well as the use of this information to feed the forecasting system. For this purpose, we have taken a different approach from that of time series forecasting, representing it as a two-stage pattern recognition problem. We have developed a peak classification model followed by a forecasting model to estimate an aggregated demand volume. We have utilized a set of machine learning algorithms to benefit from both accurate detection of the peaks and precise forecasts, as applied to the Polish power system. The key finding is that the algorithms can detect 96.3% of electricity peaks (load value equal to or above the 99th percentile of the load distribution) and deliver accurate forecasts, with mean absolute percentage error (MAPE) of 3.10% and resistant mean absolute percentage error (r-MAPE) of 2.70% for the 24 h forecasting horizon.

Artificial neural networks are capable of constructing complex decision boundaries and over the r... more Artificial neural networks are capable of constructing complex decision boundaries and over the recent years they have been widely used in many practical applications ranging from business to medical diagnosis and technical problems. A large number of error functions have been proposed in the literature to achieve a better predictive power. However, only a few works employ Tsallis statistics, which has successfully been applied in other fields. This paper undertakes the effort to examine the í µí±-generalized function based on Tsallis statistics as an alternative error measure in neural networks. The results indicate that Tsal-lis entropy error function can be successfully applied in the neural networks yielding satisfactory results.

Individual electricity customers that are connected to low voltage network in Poland are usually ... more Individual electricity customers that are connected to low voltage network in Poland are usually assigned to the most common G11 tariff group with flat prices for the whole year, no matter the usage volume. Given the diversity of customers' behavior inside the same specific group, we aim to propose an approach to assign the customers based on some objective factors rather than subjective fixed assignment. With the smart metering data and statistical methods for clustering we can explore and recommend each customer the most suitable tariff to benefit from lower prices thus generate the savings. Further, the paper applies hierarchical, k-means and Kohonen approaches to assign the customers to the proper tariff, assuming that the customer can gain the biggest expenses reduction from the tariff switch. The analysis was conducted based on the Polish dataset with an hourly energy readings among 197 entities.

Artificial neural networks are currently one of the most commonly used classifiers and over the r... more Artificial neural networks are currently one of the most commonly used classifiers and over the recent years they have been successfully used in many practical applications, including banking and finance, health and medicine, engineering and manufacturing. A large number of error functions have been proposed in the literature to achieve a better predictive power. However, only a few works employ Tsallis statistics, although the method itself has been successfully applied in other machine learning techniques. This paper undertakes the effort to examine the q-generalized function based on Tsallis statistics as an alternative error measure in neural networks. In order to validate different performance aspects of the proposed function and to enable identification of its strengths and weaknesses the extensive simulation was prepared based on the artificial benchmarking dataset. The results indicate that Tsallis entropy error function can be successfully introduced in the neural networks yielding satisfactory results and handling with class imbalance, noise in data or use of non-informative predictors.

Advanced metering infrastructures such as smart metering have begun to attract increasing attenti... more Advanced metering infrastructures such as smart metering have begun to attract increasing attention; a considerable body of research is currently focusing on load profiling and forecasting at different scales on the grid. Electricity time series clustering is an effective tool for identifying useful information in various practical applications, including the forecasting of electricity usage, which is important for providing more data to smart meters. This paper presents a comprehensive study of clustering methods for residential electricity demand profiles and further applications focused on the creation of more accurate electricity forecasts for residential customers. The contributions of this paper are threefold: (1) using data from 46 homes in Austin, Texas, the similarity measures from different time series are analyzed; (2) the optimal number of clusters for representing residential electricity use profiles is determined; and (3) an extensive load forecasting study using different segmentation-enhanced forecasting algorithms is undertaken. Finally, from the operator's perspective, the implications of the results are discussed in terms of the use of clustering methods for grouping electrical load patterns.

Peak load management allows utilities to reduce demand for electricity and optimal utilization of... more Peak load management allows utilities to reduce demand for electricity and optimal utilization of available electricity during peak usage time. Accurate peak load forecasting is crucial for such task. In this paper, we used data mining scheme to detect the peak load in the Polish electricity system. Deliberately, we undertook the approach different from time series forecasting and represented it as a classical pattern recognition problem. We used set of machine learning techniques to benefit from accurate detection of the power peaks. The results show that the algorithms can accurately detect 96.2% of the electricity peaks up to 24 hours ahead.

The main goal of this research is to discover the structure of home appliances usage patterns, he... more The main goal of this research is to discover the structure of home appliances usage patterns, hence providing more intelligence in smart metering systems by taking into
account the usage of selected home appliances and the time of their usage. In particular, we present and apply a set of unsupervised machine learning techniques to reveal specific
usage patterns observed at an individual household. The work delivers the solutions applicable in smart metering systems that might: (1) contribute to higher energy awareness; (2) support accurate usage forecasting; and (3) provide the input for demand response systems in homes with timely energy saving recommendations for users. The results provided in this paper show that determining household characteristics from smart meter data is feasible and allows for quickly grasping general trends in data.

Acta Physica Polonica A, 2015

—The Grade Correspondence Analysis (GCA) with posterior clustering and visualization is introduce... more —The Grade Correspondence Analysis (GCA) with posterior clustering and visualization is introduced and applied to individual households' electricity usage data. The main task of this analysis is to identify a way of representing the variability of a households behavior and to develop an efficient way of clustering the households into a few, usable and homogenous groups. The regularity in terms of the electricity usage is useful information for organizations to allow accurate demand planning with the aim of improving the overall efficiency of the network. The approach is tested using data from 46 households located in Austin, Texas, USA and monitored for 14 months at a sampling interval of 1 hour.

Acta Physica Polonica A, 2015