Wyniki wyszukiwania - BazTech

1

Analiza statystyczna danych oraz prognozy rynkowych cen energii (RCE) w Polsce z wyprzedzeniem do 1 godziny

Piotrowski Paweł, Baranowski Marcin, Bączek Adrian, Niedźwiecki Radosław

Elektro Info

|

2023

|

nr 4

81--85

PL

W artykule przedstawiono analizę statystyczną danych oraz prognozy rynkowych cen energii (RCE) z wyprzedzeniem do 1 godziny. Sformułowano wnioski końcowe z wykonanych prognoz oraz analiz statystycznych.

EN

The article presents a statistical analysis of data and forecasts of energy prices (RCE) in Poland up to 1 hour ahead. The conclusions have been drawn based on forecasts outcome and statistical analysis.

2

Weather-based maize yield forecast in Saudi Arabia using statistical analysis and machine learning

Jayagopal Prabhu, Muthukumaran Venkatesan, Koti Manjula Sanjay, Kumar Sampath Satheesh, Rajendran Sukumar, Mathivanan Sandeep Kumar

Acta Geophysica

|

2022

|

Vol. 70, no 6

2901--2916

EN

Crop yield is completely vulnerable to extreme weather events. Growing research investigation to establish climate change, implications in the sectors are influencing the connection. Forecasting maize output with some lead time can help producers to prepare for requirement and, in many cases, limited human resources, as well as support in strategic business decisions. The major purpose is to illustrate the relationship between various climatic characteristics and maize production, as well as to predict forecasts using ARIMA and machine learning approaches. When compared to ARIMA, the proposed method performs better in forecasting maize yields. Consequently, the neural network provides the majority of the prospective talents for forecasting maize production. Seasonal growth is susceptible of forecasting crop yields with tolerable competencies, and efforts are essential to quantify the proposed methodology that forecasts overall crop yield in diverse neighbourhoods in Saudi Arabia’s regions. The proposed combined ARIMA-LSTM model requires less training, with parameter adjustment having less effect on data prediction without bias. To monitor progress, the model may be trained repeatedly using roll back. The correlations between estimated yield and measured yield at irrigation and rain-fed sites were analysed to further validate the robustness of the optimal ARIMA-LSTM method, and the results demonstrated that the proposed model can serve as an effective approach for different types of sampling sites and has better adaptability to inter-annual fluctuations in climate with findings indicating a dependable and viable method for enhancing yield estimates.

3

Sztuczna inteligencja – klucz do rozwoju sieci 5G

Koper Zbigniew

Wiadomości Elektrotechniczne

|

2021

|

R. 89, nr 11

13--16

PL

Rozwój sieci 5G jest niemożliwy bez zastosowania sztucznej inteligencji (AI). Zastosowania AI obejmuje planowanie sieci, diagnostykę sieci oraz jej optymalizację i kontrolę. W artykule omówiono wybrane zagadnienia związane z wykorzystaniem ML, koncentrując się na szeregach czasowych i ich wykorzystaniu do przewidywania stanu sieci.

4

Matrix profile for DDoS attacks detection

Alotaibi Faisal, Lisitsa Alexei

Annals of Computer Science and Information Systems

|

2021

|

Vol. 25

357--361

EN

Previous studies have focused on DDoS, which are a crucial problem in network security. This study explore a time series method MP, which has shown effective results in a number of applications. The MP is potentially well suited to use for DDoS as a rapid method of detection,a factor that is vital for the successful identification and cessation of DDoS.The study examined how the MP performed in diverse situations related to DDoS, as well as identifying those features that are most applicable in various scenarios.Results show the efficiency of MP against all types of DDoS with the exception of NTP.

5

Impact of time series clustering on fuel sales prediction results

Henzel Joanna, Sikora Marek, Bularz Jakub

Annals of Computer Science and Information Systems

|

2021

|

Vol. 26

13--21

EN

The purpose of the paper is to check the impact of data clustering in the process of predicting demand. We checked different ways of adding information about similar datasets to the forecasting process and we grouped the measurements in multiple ways. The experiments were executed on 50 time series describing fuels sales (gasoline and diesel sales) on 25 petrol stations from an international company. We described the data preparation process and feature extraction process. In the 9 presented experiments, we used the XGBoost algorithm and some typical time series forecasting methods (ARIMA, moving average). We showed a case study for two datasets and we discussed the practical usage of the tested solutions. The results showed that the solution which used XGBoost model utilising data gathered from all available petrol stations, in general, worked the best and it outperformed more advanced approaches as well as typical time series methods.

6

Analysis of the Effect of News Sentiment on Stock Market Prices through Event Embedding

Sridhar Sashank, Sanagavarapu Sowmya

Annals of Computer Science and Information Systems

|

2021

|

Vol. 25

147--150

EN

Stock market price prediction models have remained a prominent challenge for the investors owing to their volatile nature. The impact of macroeconomic events such as news headlines is studied here using a standard dataset with closing stock price rates for a chosen period by performing sentiment analysis using a Random Forest classifier. A Bi-LSTM time-series forecasting model is constructed to predict the stock prices by using the polarity of the news headlines. It is observed that Random Forest Classifiers predict the polarity of news articles with an accuracy of 84.92%.

7

Innovative trend pivot analysis method (ITPAM): a case study for precipitation data of Susurluk Basin in Turkey

Ceribasi Gokmen, Ceyhunlu Ahmet Iyad, Ahmed Naveed

Acta Geophysica

|

2021

|

Vol. 69, no. 4

1465--1480

EN

Throughout the geological history of the earth, there have been many climate changes due to natural and external factors. In the past, the changes in climate were caused by natural causes, and today it is primarily caused by human activities. Besides being diferent climate types, Turkey is among countries that will be afected by climate change induced by global warming. Climate changes in the regions will be afected diferently and degrees due to the country’s surroundings by seas, fragmented topography and orographic features. Trend analysis methods are used in many areas such as on various engi neering, agriculture, environmental and water resources, especially in climate change impact studies resulting from global warming. When data are analyzed with classical trend analysis methods, forward-looking predictions are generally made as low, medium, high, decreasing and increasing. However, risk classes showing changes between available data sets are not known. Innovative Trend Pivot Analysis Method (ITPAM) determines risk classes by establishing a relationship between data. Furthermore, in this method, increasing and decreasing trend regions are separated into fve classes more clearly than classical/traditional trend methods. In this study, Susurluk Basin’s total monthly precipitation data (2006–2017) were analyzed by using ITPAM which the newest trend method. When arithmetic mean analysis results are examined, a signifcant change is observed between frst data set and second data set at two stations (Bandirma and Uludag). When examined at other stations, it is observed that at least one month of almost every station is in 1st degree risk group. When standard deviation analysis results of each station are examined, a signifcant change is observed between frst data set and second data set at many stations. Because while trend class of a point in developed IPTA graph is the medium degree, this point is in 1st risk class in the risk graph.

8

Research on the combustion process using time series

Grądz Żaklin

Informatyka, Automatyka, Pomiary w Gospodarce i Ochronie Środowiska

|

2020

|

T. 10, nr 2

52--55

EN

In the combustion process, one of the most important tasks is related to maintaining its stability. Numerous methods of monitoring, diagnostics, and analysis of the measurement data are used for this purpose. The information recorded in the combustion chamber constitute one-dimensional time series. In the case of non-stationary time series, which can be transformed into the stationary form, the autoregressive integrated moving average process can be employed. The paper presented the issue of forecasting the changes in flame luminosity. The investigations discussed in the work were carried out with the ARIMA model (p,d,q). The presented forecasts of changes in flame luminosity reflect the actual processes, which enables to employ them in diagnostics and control of the combustion process.

PL

W procesie spalania jednym z najważniejszych zadań jest zachowanie jego stabilności. Do tego celu wykorzystywanych jest wiele metod z zakresu monitorowania, diagnostyki i analizy danych pomiarowych. Zarejestrowane w komorze spalania informacje są jednowymiarowymi szeregami czasowymi. W przypadku niestacjonarnych szeregów czasowych, które można przekształcić do formy stacjonarnej, znalazły zastosowanie scałkowane procesy autoregresji i średniej ruchomej. W artykule przedstawiono problematykę prognozowania zmian intensywności świecenia płomienia. Badania zaprezentowane w pracy zostały przeprowadzone z wykorzystaniem modelu ARIMA(p,d,q). Przedstawione prognozy zmian intensywność świecenia płomienia odwzorowują rzeczywiste przebiegi, co pozwala wykorzystać je w diagnostyce i sterowaniu procesem spalania.

9

Analiza rozwoju elektromobilności w Polsce oraz prognozy liczby pojazdów z napędem elektrycznym do roku 2025

Piotrowski Paweł

Elektro Info

|

2020

|

nr 5

74--77

PL

Analiza rozwoju elektromobilności w Polsce oraz prognozy liczby pojazdów z napędem elektrycznym do roku 2025

EN

The analysis of e-mobility development in Poland and forecasts of the number electric vehicles by 2025

10

Adaptation of models from determined chaos theory to short-term power forecasts for wind farms

Popławski T., Szeląg P., Bartnik R.

Bulletin of the Polish Academy of Sciences. Technical Sciences

|

2020

|

Vol. 68, nr 6

1491--1501

EN

The paper proposes an adaptation of mathematical models derived from the theory of deterministic chaos to short-term power forecasts of wind turbines. The operation of wind power plants and the generated power depend mainly on the wind speed at a given location. It is a stochastic process dependent on many factors and very difficult to predict. Classical forecasting models are often unable to find the existing relationships between the factors influencing wind power output. Therefore, we decided to refer to fractal geometry. Two models based on self-similar processes (M-CO) and (M-COP) and the (M-HUR) model were built. The accuracy of these models was compared with other short-term forecasting models. The modified model of power curve adjusted to local conditions (M-PC) and Canonical Distribution of the Vector of Random Variables Model (CDVRM). Examples of applications confirm the valuable properties of the proposed approaches.

11

Water quality management in a treatment plant using the Box-Jenkins method

Redondo-Peñuela E. A., Rondón-Quintana H. A., Zafra-Mejía C. A.

Archives of Civil Engineering

|

2020

|

Vol. 66, nr 3

125--137

EN

Drinking water systems are critical to society. They protect residents from waterborne illnesses and encourage economic success of businesses by providing consistent water supplies to industries and supporting a healthy work force. This paper shows a study on water quality management in a treatment plant (TP) using the Box-Jenkins method. A comparative analysis was carried out between concentrations of water quality parameters, and Colombian legislation and guidelines established by the World Health Organization. We also studied the rainfall influence in relation to variations in water quality supplied by the TP. A correlation analysis between water quality parameters was carried out to identify management parameters during the TP operation. Results showed the usefulness of the Box-Jenkins method for analyzing the TP operation from a weekly timescale (mediumterm), and not from a daily timescale (short-term). This was probably due to significant daily variations in the management parameters of water quality in the TP. The application of a weekly moving average transformation to the daily time series of water quality parameter concentrations significantly decreased the mean absolute percentage error in the forecasts of Box-Jenkins models developed. Box-Jenkins analysis suggested an influence of the water quality parameter concentrations observed in the TP during previous weeks (between 2-3 weeks). This study was probably constituted as a medium-term planning tool in relation to atypical events or contingencies observed during the TP operation. Finally, the findings in this study will be useful for companies or designers of drinking water treatment systems to take operational decisions within the public health framework.

12

Generating Fuzzy Linguistic Summaries for Menstrual Cycles

Sosnowski Łukasz, Penza Tomasz

Annals of Computer Science and Information Systems

|

2020

|

Vol. 21

119--128

EN

This paper presents a method of generating linguistic summaries of women's menstrual cycles based on the set of concepts describing various aspects of the cycles. These concepts enable description of menstrual cycles that are readable for humans, but they also provide high-level information that can be used as control input for other data processing actions such as e.g. anomaly detection. The labels signifying these concepts are assigned to cycles by means of multivariate time series analysis. The corresponding algorithm is a subsystem of a bigger solution created as a part of an R&D project.

13

Deep Bi-Directional LSTM Networks for Device Workload Forecasting

Ruta Dymitr, Cen Ling, Vu Quang Hieu

Annals of Computer Science and Information Systems

|

2020

|

Vol. 21

115--118

EN

Deep convolutional neural networks revolutionized the area of automated objects detection from images. Can the same be achieved in the domain of time series forecasting? Can one build a universal deep network that once trained on the past would be able to deliver accurate predictions reaching deep into the future for any even most diverse time series? This work is a first step in an attempt to address such a challenge in the context of a FEDCSIS'2020 Competition dedicated to network device workload prediction based on their historical time series data. We have developed and pre-trained a universal 3-layer bi-directional Long-Short-Term-Memory (LSTM) regression network that reported the most accurate hourly predictions of the weekly workload time series from the thousands of different network devices with diverse shape and seasonality profiles. We will also show how intuitive human-led post-processing of the raw LSTM predictions could easily destroy the generalization abilities of such prediction model.

14

A Framework for Time Series Preprocessing and History-based Forecasting Method Recommendation

Züfle Marwin, Kounev Samuel

Annals of Computer Science and Information Systems

|

2020

|

Vol. 21

141--144

EN

The complexity of managing the capacities of large IT infrastructures is constantly increasing as more network devices are connected. This task can no longer be performed manually, so the system must be monitored at runtime and estimations of future conditions must be made automatically. However, since using a single forecasting method typically performs poorly, this paper presents a framework for forecasting univariate network device workload traces using multiple forecasting methods. First, the time series are preprocessed by imputing missing data and removing anomalies. Then, different features are derived from the univariate time series, depending on the type of forecasting method. In addition, a recommendation approach for selecting the most suitable forecasting method from this set of algorithms for each time series based only on its historical values is proposed. For this purpose, the performance of the forecasting methods is approximated using the historical data of the respective time series under consideration. The framework is used in the FedCSIS 2020 Challenge and shows good forecasting quality with an average R2 score of 0.2575 on the small test data set.

15

Relationship between selected percentiles and return periods of extreme events

Camufo Dario, Becherini Francesca, della Valle Antonio

Acta Geophysica

|

2020

|

Vol. 68, no. 4

1201--1211

EN

This paper investigates the relationship between selected percentiles, return periods and the concepts of rare and extreme events in climate and hydrological series, considering both regular and irregular datasets, and discusses the IPCC and WMO indications. IPCC (Annex II: Glossary. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. IPCC, Geneva, 2014) establishes that an extreme event should be rare and exceed selected upper and lower thresholds (10th and 90th percentiles); WMO (Guidelines on the defnition and monitoring of extreme weather and climate events-TT-DEWCE WMO 4/14/2016. World Meteorological Organization, Geneva, 2016) suggests thresholds near the ends of the range, but leaves them undetermined. The concept of “rare” relates the extreme events to the time domain and is typically expressed in terms of return period (RP). The key is to fnd the combination between “rare”, percentile and return period. In particular, two crucial items are analysed: (1) how the return period may vary in response to the choice of the threshold, in particular when it is expressed in terms of percentiles; (2) how the choice of producing a regular or irregular dataset may afect the yearly frequency and the related return periods. Some weather variables (e.g. temperature) are regular and recorded at fxed time intervals, while other phenomena (e.g. tornadoes) occur at times. Precipitation may be considered either regular, all-days being characterized by a precipitation amount from 0 (no precipitation) to the top of the range, or irregular (rainy-days only) considering a precipitation day over a selected instrumental or percentile threshold. These two modes of interpreting precipitation include a diferent number of events per year (365 or less) and generate diferent return periods. Every climatic information may be afected by this defnition. The 90th percentile applied to observations with daily frequency produces 10-day return period and the percentiles necessary to get 1 year, 10 years or other return periods are calculated. The general case of events with selected or variable frequencies, and selected percentiles, is also considered with an example of a precipitation series, two-century long.

16

Identyfikacja profilu osobowego na podstawie analizy neuronowej parametrów środowiskowych w inteligentnym budynku

Stachno Andrzej, Suproniuk Marek, Skibko Zbigniew

Przegląd Elektrotechniczny

|

2019

|

R. 95, nr 11

133--135

PL

Dla potrzeb takiej identyfikacji osób przebywających w pomieszczeniach budynu, opracowany został algorytm profilowania i identyfikacji osobowej z zastosowaniem sztucznych sieci neuronowych – neuronową identyfikacją organicznego profilu osobowego (NIOPO – ang. Neural identification of an organic personal profile NIOPP). Identyfikacja neuronowa, wykorzystuje pomiary koncentracji gazów, których proporcje oraz skład są cechą indywidualną dla każdego człowieka.

EN

For the purpose of such identification of people staying in the building's premises, a profiling and personal identification algorithm was developed with the use of artificial neural networks - neural identification of the organic personal profile (NIOPP). Neural identification, uses measurements of gas concentrations whose proportions and composition are an individual feature for every human being.

17

Times series averaging and denoising from a probabilistic perspective on time-elastic kernels

Marteau Pierre-Francois

International Journal of Applied Mathematics and Computer Science

|

2019

|

Vol. 29, no. 2

375--392

EN

In the light of regularized dynamic time warping kernels, this paper re-considers the concept of a time elastic centroid for a set of time series. We derive a new algorithm based on a probabilistic interpretation of kernel alignment matrices. This algorithm expresses the averaging process in terms of stochastic alignment automata. It uses an iterative agglomerative heuristic method for averaging the aligned samples, while also averaging the times of their occurrence. By comparing classification accuracies for 45 heterogeneous time series data sets obtained by first nearest centroid/medoid classifiers, we show that (i) centroid-based approaches significantly outperform medoid-based ones, (ii) for the data sets considered, our algorithm, which combines averaging in the sample space and along the time axes, emerges as the most significantly robust model for time-elastic averaging with a promising noise reduction capability. We also demonstrate its benefit in an isolated gesture recognition experiment and its ability to significantly reduce the size of training instance sets. Finally, we highlight its denoising capability using demonstrative synthetic data. Specifically, we show that it is possible to retrieve, from few noisy instances, a signal whose components are scattered in a wide spectral band.

18

An algorithm for arbitrary-order cumulant tensor calculation in a sliding window of data streams

Domino Krzysztof, Gawron Piotr

International Journal of Applied Mathematics and Computer Science

|

2019

|

Vol. 29, no. 1

195--206

EN

High-order cumulant tensors carry information about statistics of non-normally distributed multivariate data. In this work we present a new efficient algorithm for calculation of cumulants of arbitrary orders in a sliding window for data streams. We show that this algorithm offers substantial speedups of cumulant updates compared with the current solutions. The proposed algorithm can be used for processing on-line high-frequency multivariate data and can find applications, e.g., in on-line signal filtering and classification of data streams. To present an application of this algorithm, we propose an estimator of non-Gaussianity of a data stream based on the norms of high order cumulant tensors. We show how to detect the transition from Gaussian distributed data to non-Gaussian ones in a data stream. In order to achieve high implementation efficiency of operations on super-symmetric tensors, such as cumulant tensors, we employ a block structure to store and calculate only one hyper-pyramid part of such tensors.

19

Method of diagnostics of filling material strength based on time series

Mustafin Salim, Arslanov Marat, Zeinullin Abdikarim, Korobova Ekaterina

Informatyka, Automatyka, Pomiary w Gospodarce i Ochronie Środowiska

|

2019

|

T. 9, nr 3

46--49

EN

The method of forecasting of the filling concreting process according to the observational data of the stowing material’s state indices is carried out at earlier time series. The method of predicting the hardening process is based on pattern recognition methods. An algorithm for the case when the training set contains sets of time series of several classes is proposed.

PL

W pracy omówiono metodę prognozowania wypełnień w procesie betonowania, na podstawie danych obserwacyjnych stanu materiału sztauerskiego, w oparciu o zarejestrowane uprzednio szeregi czasowe. Zaproponowane rozwiązanie dotyczące procesu utwardzania bazuje na metodach rozpoznawania wzorców. Zaproponowano algorytm dla przypadku, gdy zestaw uczący zawiera zestawy szeregów czasowych dla kilku klas.

20

Metodyka wyboru metody prognozowania wizyt na stronie internetowej w przedsiębiorstwie turystycznym

Kozicki Bartosz, Tomaszewski Jarosław

Gospodarka Materiałowa i Logistyka

|

2019

|

nr 7

20--43

PL

W artykule poruszono problem badawczy związany z analizą danych pierwotnych, jej oceną, wyborem najlepszej metody do prognozy na przyszłość. Badania rozpoczęto od wykorzystania narzędzi badawczych do poszukiwania istnienia w szeregu czasowym pierwotnym zależności. Następnym krokiem było pogrupowanie danych, ich analiza i ocena. Uzyskane oceny stały się przesłanką zbudowania modelu zerojedynkowego regresji wielorakiej w celu potwierdzenia wykrytych zależności. Stwierdzone zależności pozwoliły na sprowadzenie szeregu pierwotnego do stacjonarności. Szereg czasowy został podzielony na dwie części: uczącą i testową. Wskutek krytycznej analizy literatury i uzyskanych zależności wybrano trzy najlepsze metody do prognozy szeregu uczącego na okres równy szeregowi testowemu. Uzyskane prognozy zostały poddane ocenie przy zastosowaniu obserwacji wzrokowej i MAPE. Wybrano najlepszą metodę, którą wykonano prognozowanie szeregu czasowego pierwotnego na 2019 rok (202 przyszłe okresy).

EN

In this article the author raises the research problem regarding the analysis of original data, its evaluation and the selection of the best forecasting method for the future. The research was initiated with the application of research tools in order to search for the relationships within the original time series. The following step was to group data, analyze and evaluate them. The results obtained were the premise for the construction of a zero-one model of multiple regression in order to confirm the relationships found. The detected relationships enabled to bring down the original series to stationarity. Time series was divided into two parts: teaching and testing ones. Due to the critical analysis of literature and the relationships obtained, three best forecasting methods of teaching series were selected for the same period of testing series. The forecasting obtained were evaluated by means of visual observation and MAPE. The best method was selected for the forecasting of the original time series for 2019 (202 future periods).