Wyniki wyszukiwania - BazTech

1

High-Resolution Lithology Detection Using Sentinel-2A, ALOS PRISM L1B Images, and Support-Vector Machines in Tagragra d’Akka Inlier of Western Anti-Atlas, Morocco

Hammoud Yassine, Allali Youssef, Saadane Abderrahim

Geomatics and Environmental Engineering

|

2025

|

Vol. 19, no. 1

103--141

EN

Geological mapping faces substantial challenges due to inaccessible terrains, labor-intensive field methods, and potential interpretative errors. This study proposes an innovative approach that leverages automatic lithology classification using multispectral Sentinel-2A (10 m) and high-resolution panchromatic ALOS PRISM L1B (2.5 m) images. Applied to the Tagragra d’Akka inlier of the Anti-Atlas region, the methodology enhances spatial resolution through pansharpening, followed by unsupervised segmentation. The segmented images are classified using support vector machines (SVMs) (supervised learning algorithms) to distinguish the lithological units. Achieving an 86% overall accuracy and an 84% kappa coefficient, the approach demonstrated robust performance and surpassed conventional techniques. The integration of machine learning and remote sensing offers a promising frontier for geological mapping – particularly in regions like the Tagragra d’Akka inlier. This study marks a significant advancement in automating lithological mapping, with implications for geological research, resource management, and hazard assessment. Automated techniques in geological cartography significantly enhance mapping accuracy and efficiency. Future studies should explore additional data sources and machine-learning algorithms to refine lithological classification and validate these methods across diverse geological settings.

2

Power transformer fault warning combining support vector machine and improved grey wolf optimization algorithm

Zhao Shuzong, Pattanadech Norasage

Archives of Electrical Engineering

|

2025

|

Vol. 74, nr 1

191--208

EN

To optimize the parameter setting of the support vector machine and improve the classification performance and computational efficiency of power transformer fault diagnosis, this study proposes an improved grey wolf optimization algorithm. By optimizing the global search and local optimization capabilities of the grey wolf algorithm and combining them with stacked denoising autoencoders, a new power transformer fault warning model is constructed. Firstly, the grey wolf optimization algorithm is optimized through four strategies: elite reverse learning, nonlinear control parameters, Lévy flight, and particle swarm optimization, which improve its global search and local optimization capabilities. Secondly, the stacked denoising autoencoder is utilized to extract high-level features of fault data, and the improved GWO algorithm and SVM are combined to complete fault classification. The results indicated that the proposed diagnostic model achieved a diagnostic accuracy of 0.979, a recall rate of 0.986, and an F1 value of 0.983 in benchmark performance testing. In practical applications, the average fault diagnosis accuracy of this model could reach up to 99.21%, and the average diagnosis time was only 0.08 s. The developed power transformer fault warning model can provide an efficient and reliable technical solution for fault diagnosis in the power system.

3

Detecting clustered fruits using a hybrid of convolutional neural networks and machine learning classifiers – Case study

Falih Bashar S., Gierz Łukasz, Al-Zaidi Ghazwan A.

Advances in Science and Technology. Research Journal

|

2025

|

Vol. 19, no. 4

1--9

EN

The last and most important procedure during fruit or vegetable cultivation is harvesting. One of the basic challenges during grape growing is the use of agriculture 4.0 machines (including robots) during harvesting which is associated with the need for quick identification of berries or grape clusters. In this work, a convolutional neural network (CNN) and a machine learning classifier were suggested for the identification (detection) of individual grapes. A free data set (Iceland) was used, which included two classes with different lighting conditions and berry sizes. The integrated method included two types of deep learning models, i.e. CNN (AlexNet and GoogleNet). CNN models were used to obtain discriminative deep features from different layers. The combination of two models AlexNet-Fc6 and SVM-Cubic yielded the highest accuracy, sensitivity and precision (mean ± standard deviation) % of 99.4 ± 0.13, 99.2 ± 0.14 and 99.49 ± 0.19, respectively. The developed grape detector can be used for practical applications requiring high accuracy, e.g. in the process of yield estimation or detection of grape diseases.

4

Acoustic model for the classification of Polish vowels

Pondel-Sycz Karolina

Vibrations in Physical Systems

|

2024

|

Vol. 35, nr 1

art. no. 2024101

EN

The study explored the performance of vowel recognition using an acoustic model built on Audio Fingerprint techniques [1]. The research compares the performance of Support Vector Machines (SVMs), Hidden Markov Models (HMMs), Artificial Neural Networks (ANNs) and k-Nearest Neighbours (k-NN) classifiers in the recognition of isolated and within-word vowels and investigates the importance of different types of acoustic speech features in this process. Temporal, spectral, cepstral, formant, LPC and perceptual features of speech were examined. Importance of features was tested using a random forest classifier. Vowel classification was tested at three confidence levels for feature importance: 90%, 95% and 99%. Two author databases consisting of a total of 1,200 samples from 20 speakers, recorded under household conditions, were used. The classifiers were evaluated by confusion matrix, accuracy, precision, sensitivity and F1 score. A segmentation of words into speech sounds was carried out using a tool based on BiLSTM recurrent neural networks and the BIC criterion. Three most important features were determined: power spectral density, spectral cut-off, and Power-Normalised Cepstral Coefficients. In the isolated vowel recognition task, the SVM classifier was the most effective with a feature significance confidence level of 95% obtaining accuracy = 81%, precision = 81%, sensitivity = 81%, F1 score = 80%. In the task of recognising a vowel within a word, it was verified if the algorithm detected the presence of vowels in the correct segment and if it recognised the correct vowel within it. The best results were obtained by the k-NN classifier (statistical confidence level of feature importance of 99.9%). However, these results were low, correct recognition of the vowel in the word: A, E, U: 20%, I, O: 7%, Y: 23%. This indicates strong influence of the neighbourhood of other speech sounds in speech on the acoustic model of vowels and their recognition.

5

Analysis of the Effectiveness of Selected Machine Learning Algorithms in the Classification of Satellite Image Content Depending on the Size of the Training Sample

Kupidura Przemysław, Niemyski Stanisław

Teledetekcja Środowiska

|

2024

|

T. 64

24--38

EN

The article presents an analysis of the accuracy of 3 popular machine learning (ML) methods: Maximum Likelihood Classifier (MLC), Support Vector Machine (SVM), and Random Forest (RF) depending on the size of the training sample. The analysis involved performing the classification of the content of a Landsat 8 satellite image (divided into 6 basic land cover classes) in 10 different variants of the number of training samples (from 2664 to 34711 pixels), estimating individual results, and a comparative analysis of the obtained results. For each classification variant, an error matrix was developed and on their basis, accuracy metrics were calculated: f1-score, precision and recall (for individual classes) as well as overall accuracy and kappa index of agreement (generally for the entire classification). The analysis showed a stimulating effect of the size of the training sample on the accuracy of the obtained classification results in all analyzed cases, with the most sensitive to this factor being MLC, showing the best effectiveness with the largest training sample and the smallest - with the smallest, and the least SVM, characterized by the highest accuracy with the smallest training sample, comparing to other algorithms.

PL

Artykuł przedstawia analizę dokładności 3 popularnych metod uczenia maszynowego: Maximum Likelihood Classifier (MLC), Support Vector Machine (SVM) oraz Random Forest (RF) w zależności od liczebności próbki treningowej. Analiza polegała na wykonaniu klasyfikacji treści zdjęcia satelitarnego Landsat 8 (w podziale na 6 podstawowych klas pokrycia terenu) w 10 różnych wariantach liczebności próbek uczących (od 2664 do 34711 pikseli), oszacowaniu poszczególnych wyników oraz analizie porównawczej uzyskanych wyników. Dla każdego wariantu klasyfikacji opracowano macierz błędów, a na ich podstawie obliczono metryki dokładności: F1-score, precision and recall (dla pojedynczych klas) oraz ogólną dokładność i wskaźnik zgodności Kappa (ogólnie dla całej klasyfikacji). Analiza wykazała stymulujący wpływ rozmiaru próbki uczącej na dokładność uzyskiwanych wyników klasyfikacji we wszystkich analizowanych przypadkach, przy czym najbardziej wrażliwym na ten czynnik był MLC, wykazujący się najlepszą skutecznością przy największej próbce treningowej i najmniejszą - przy najmniejszej, a najmniej SVM, cechujący się największą dokładnością przy najmniejszej próbce treningowej, w porównaniu do pozostałych algorytmów.

6

MPPT of PEM fuel cell using DC-DC boost converter based on SVM

Agwa Ahmed M., Alruwaili Mohammed

Przegląd Elektrotechniczny

|

2024

|

R. 100, nr 11

53--58

EN

Detrimental environmental influences and restricted quantities of conventional energies impose the employment of renewable energies (REs). Regrettably, REs for instance wind and solar energies are sporadic, therefore they have to be stored in different forms for employment throughout their absenteeism. For that purpose, REs can be stored excellently through generation of hydrogen using electrolyzer throughout abundance, then generation of electricity using fuel cell (FC) throughout their absenteeism. Concerning the merits of the proton exchange membrane FC (PEMFC), it is recommended more than different types of FCs. The PEMFC power lacks constancy, as it relies on pressure of hydrogen, temperature, and loading. Hence, a maximum power point tracking (MPPT) technique have to be employed with PEMFC. The procedures formerly employed possess some demerits, for instance delay of reaction, immensity of oscillation, and hugeness of overshoot and undershoot, accordingly this research addresses a PEMFC MPPT based on support vector machine (SVM). Simulation findings of employing the SVM for PEMFC MPPT expose its merits over other techniques in terms of equilibrium among speediness of reaction, tininess of oscillations, and smallness of overshoot and undershoot.

PL

Niekorzystne wpływy środowiskowe i ograniczone ilości konwencjonalnych energii wymuszają wykorzystanie odnawialnych źródeł energii (RE). Niestety, RE, na przykład energia wiatrowa i słoneczna, są sporadyczne, dlatego muszą być przechowywane w różnych formach do wykorzystania w czasie nieobecności. W tym celu RE mogą być doskonale przechowywane poprzez wytwarzanie wodoru za pomocą elektrolizera w obfitości, a następnie wytwarzanie energii elektrycznej za pomocą ogniwa paliwowego (FC) w czasie nieobecności. Jeśli chodzi o zalety membrany wymiany protonów FC (PEMFC), jest ona bardziej zalecana niż inne typy FC. Moc PEMFC nie jest stała, ponieważ opiera się na ciśnieniu wodoru, temperaturze i obciążeniu. Dlatego też w przypadku PEMFC należy zastosować technikę śledzenia maksymalnego punktu mocy (MPPT). Wcześniej stosowane procedury mają pewne wady, na przykład opóźnienie reakcji, ogrom oscylacji i ogrom przekroczenia i niedoregulowania, dlatego też niniejsze badania dotyczą MPPT PEMFC opartego na maszynie wektorów nośnych (SVM). Wyniki symulacji wykorzystującej SVM do pomiaru MPPT PEMFC ujawniają jej zalety w porównaniu z innymi technikami w zakresie równowagi między szybkością reakcji, niewielkimi oscylacjami oraz niewielkimi przekroczeniami i niedoregulowaniami.

7

Induction motor failures detection using motor current signal analysis (MCSA) and two-step support vector machine (SVM) classifier

Ghediri Abdelhalim, Goléa Noureddine

Przegląd Elektrotechniczny

|

2024

|

R. 100, nr 2

272--278

EN

The induction motor, because of its robustness, is very widespread in the industrial environment. Its use is firstly at the training of nonvariable speed systems. Today, with the improvement of the power electronics, the asynchronous motor can be servo-controlled both in torque and speed. The asynchronous machine is no longer limited to constant speed applications. It is therefore found more and more in electric traction, but also in electromechanical actuators which require a control in position and for which this type of engine was not previously usable, but it is not immune to several anomalies in the industrial field, such as breaks in bars, short circuits, misalignment. So then this article enters the diagnostic framework, through the use of the SVM formalism for the localization of electrical faults in induction motor. This study is based on the stepwise application of the SVM, firstly, the nature of the fault (break of bar or short circuit of turns in the phases of the stator) is compared with the normal behavior of the asynchronous machine, this step requires the use of the SVM one against one (One vs. One). Secondly we consider the fault detected by the first analysis and try to rank it and classify it among different classes of faults of the same nature, which can give a degree of severity a reliable decision to maintenance expertise. It uses the technique of one against all (One vs. All). The proposed approach is based on the use of feature extraction especially the amplitudes and frequencies, reflecting the behavior of the induction motor, using the Motor Current Signal Analysis (MCSA); then the classification of these characteristics is realized by SVM method. The SVM classification is conducted on a 1Kw induction machine experimental benchmark with different faulty operating conditions.

PL

Silnik indukcyjny, ze względu na swoją wytrzymałość, jest bardzo rozpowszechniony w środowisku przemysłowym. Jego zastosowanie to przede wszystkim trening systemów bez zmiennej prędkości. Obecnie, wraz z udoskonaleniem elektroniki mocy, silnik asynchroniczny może być sterowany serwomechanizmem zarówno pod względem momentu obrotowego, jak i prędkości. Maszyna asynchroniczna nie jest już ograniczona do zastosowań o stałej prędkości. Dlatego coraz częściej znajduje się w trakcji elektrycznej, ale także w siłownikach elektromechanicznych, które wymagają kontroli położenia i dla których ten typ silnika nie był wcześniej używany, ale nie jest odporny na kilka anomalii w dziedzinie przemysłu, takich jak pęknięcia prętów, zwarcia, niewspółosiowość. Tak więc ten artykuł wchodzi w ramy diagnostyczne, poprzez wykorzystanie formalizmu SVM do lokalizacji uszkodzeń elektrycznych w silniku indukcyjnym. Niniejsze badanie opiera się na stopniowym stosowaniu SVM, po pierwsze, charakter usterki (przerwanie szyny lub zwarcie zwojów w fazach stojana) jest porównywany z normalnym zachowaniem maszyny asynchronicznej, ten krok wymaga użycia SVM jeden na jednego (jeden na jednego). Po drugie, rozważamy usterkę wykrytą podczas pierwszej analizy i próbujemy ją uszeregować i sklasyfikować wśród różnych klas usterek o tym samym charakterze, co może dać pewien stopień dotkliwości ekspertom w zakresie konserwacji. Wykorzystuje technikę jeden przeciwko wszystkim (jeden przeciwko wszystkim). Proponowane podejście opiera się na wykorzystaniu ekstrakcji cech, zwłaszcza amplitud i częstotliwości, odzwierciedlających zachowanie silnika indukcyjnego, z wykorzystaniem analizy sygnatury prądu silnika (MCSA); wówczas klasyfikacja tych cech jest realizowana metodą SVM. Klasyfikacja SVM jest przeprowadzana na doświadczalnym wzorcu porównawczym maszyny indukcyjnej o mocy 1 kW z różnymi wadliwymi warunkami pracy

8

Technological challenges in manufacturing of vacuum gauge thermionic cathode using thick-film technology

Jasińska Laura, Dzbik Krzysztof, Nowak Damian, Stojek Krzysztof, Chudzyńska Aleksandra, Politański Kamil, Malecha Karol

Materials Science Poland

|

2024

|

Vol. 42, No. 1

126--139

EN

This paper focuses on the development of a technological challenges of manufacturing the planar ceramic vacuum sensor based on the principles of hot-cathode ionization in the Bayard-Alpert configuration. The goal is to simplify the technological process by utilizing planar platinum structures with gold electrical paths instead of 3-dimensional structures. Various methods were tested, including the use of carbon-based SVM (Sacrifice Volume Materials) materials, but without success. Wet-etching using potassium hydroxide on Al2O3 substrates showed promise results. The findings highlight the challenges and progress made in developing the thermo-emittercomponent of the vacuum sensor.

9

Detection of human finger joints in ultrasound images : structure and optimization

Bąk Artur, Wereszczyński Kamil, Segen Jakub, Mielnik Paweł, Fojcik Marcin, Kulbacki Marek

International Journal of Electronics and Telecommunications

|

2024

|

Vol. 70, No. 2

271--276

EN

Synovitis is the inflammation of a synovial membrane surrounding a joint. Its assessment is an important step in the diagnosis and treatment of rheumatoid arthritis. Joint detection is the first stage of an automated method of assessment of a degree of synovitis, from an Ultrasound (USG) image of a finger joint and its surrounding area. A joint detector consists of three parts: image preprocessing, feature extraction, and classification. Each part contains adjustable parameters that must be set experimentally to ensure the proper operation of the detector. Both the structure of a joint detector and a procedure for finding a near-optimal configuration of the adjustable parameters are described. The optimization process is based on two evaluation measures: Area Under the Receiver Operating Characteristic Curve (AUC) and False Positive Count (FPC). The optimization process decreases the number of pictures with multiple detections, which was the main point of works presented in this paper. This was achieved by increasing the number of components of the homogeneous mixed-SURF descriptor which has the greatest influence on the final result. Non-SURF descriptors achieve poorer classification results. Our research led to the creation of a better joint detector which could positively influence the final results of inflammation level classification.

10

A zero-voltage-switching current source inverter with three-stage space vector modulation for low-input voltage stand-alone applications

Fu Xiaofeng, Zhang Mingkang, Ma Dazhuang, Jiang Jiahui, Cai Fenghuang

Archives of Electrical Engineering

|

2024

|

Vol. 73, nr 4

941--959

EN

In this paper, a three-stage space vector pulse width modulation (SVPWM) along with soft switching is proposed for a high-voltage transfer ratio single-stage three phase current-source inverter (HVTR-CSI) to reduce switching losses and improve inverter efficiency. The proposed SVPWM strategy utilizes the conduction state of the energy storage switch as the zero vector and assigning effective vectors action modes. The zero-voltage-switching (ZVS) of the energy storage switch is achieved by resonant parameters and controlling the turn-on time of the active clamp circuit. The circuit topology, operation principle, high-frequency switching process of the studied CSI are thoroughly analyzed, and detailed calculations of the circuit parameters and soft-switching design are performed. Experimental results on a 1 kW 24 VDC/84 VAC 3-phase AC prototype show that the modulation improves the CSI peak efficiency by 1.15% compared to sinusoidal pulse width modulation (SPWM). This study provides an effective design approach for the HVTR-CSI in terms of reducing switching losses.

11

The Use of Artificial Intelligence for Quality Assessment of Refill Friction Stir Spot Welded Thin Joints

Kubit Andrzej, Kłosowski Grzegorz, Berezowski Wojciech

Advances in Science and Technology. Research Journal

|

2024

|

Vol. 18, no. 3

45--57

EN

This paper presents a machine learning and image segmentation based advanced quality assessment technique for thin Refill Friction Stir Spot Welded (RFSSW) joints. In particular, the research focuses on developing a predictive support vector machines (SVM) model. The purpose of this model is to facilitate the selection of RFSSW process parameters in order to increase the shear load capacity of joints. In addition, an improved weld quality assessment algorithm based on optical analysis was developed. The research methodology includes specimen preparation stages, mechanical tests, and algorithmic analysis, culminating in a machine learning model trained on experimental data. The results demonstrate the effectiveness of the model in selecting welding process parameters and assessing weld quality, offering significant improvements compared to standard techniques. This research not only proposes a novel approach to optimizing welding parameters but also facilitates automatic quality assessment, potentially revolutionizing and spreading the application of the RFSSW technique in various industries

12

Outlier detection in EEG signals

Duraj Agnieszka, Chomątek Łukasz

Przegląd Elektrotechniczny

|

2023

|

R. 99, nr 1

237--240

EN

In this paper, the topic of detection of outliers in EEG signals was discussed, which facilitates making decisions about the diagnosis of a patient based on this study. We used two methods to detect outliers: the support vector machine and the k nearest neighbors method. The experiments were performed on a publicly available dataset containing EEG test results for 500 patients. The obtained results showed that the methods we used allow for the outlier detection efficiency at the level of 93%.

PL

W niniejszej pracy podjęto temat detekcji wyjątków w sygnałach EEG, co pozwala na ułatwienie podejmowania decyzji co do diagnozy pacjenta na podstawie tego badania. Do detekcji wyjątków wykorzystaliśmy dwie metody: maszynę wektorów nośnych i metodę k najblizszych sąsiadów. Eksperymenty zostały przeprowadzone na ogólnodostępnym zbiorze danych zawieraj ącym wyniki badania EEG dla 500 pacjentów. Uzyskane wyniki pokazały, że u żyte przez nas metody pozwalają na uzyskanie skuteczności detekcji wyjątków na poziomie 93%.

13

Machine learning to diagnose breast cancer

Ismaeel Ghassan Ahmad

Przegląd Elektrotechniczny

|

2023

|

R. 99, nr 1

10--12

EN

As the number of breast cancer diseases is increasing rapidly every year, new technologies are utilized to predict and diagnose this disease for better women's lives worldwide. The development of Machine Learning can be utilized to contribute in this sense and help in the early diagnosis of breast cancer. This paper aims to predict and diagnose breast cancer using Machine Learning techniques such as support vector Machine (SVM) and Decision -tree and Nearest neighbour (KNN). The results show the out performance of SVM over the other methods. These methods can be very helpful to predict the breast cancer disease ahead of time.

PL

Ponieważ liczba zachorowań na raka piersi gwałtownie rośnie z roku na rok, nowe technologie są wykorzystywane do przewidywania i diagnozowania tej choroby w celu poprawy życia kobiet na całym świecie. Rozwój uczenia maszynowego może być wykorzystany do wniesienia wkładu w tym sensie i pomocy we wczesnej diagnozie raka piersi. Niniejszy artykuł ma na celu przewidywanie i diagnozowanie raka piersi przy użyciu technik uczenia maszynowego, takich jak maszyna wektora nośnego (SVM) oraz drzewo decyzyjne i najbliższy sąsiad (KNN). Wyniki pokazują wydajność SVM w porównaniu z innymi metodami. Metody te mogą być bardzo pomocne w przewidywaniu zgonów na raka piersi z wyprzedzeniem.

14

Continuous speech commands recognition with thai language used support vector machine technique: A case study of speech commands control for mobile robots

Muangjaroen Supavit, Udomsiri Sakol

Przegląd Elektrotechniczny

|

2023

|

R. 99, nr 6

181--187

EN

This article presents the application of an automatic speech recognition by continuous speech commands recognition with Thai language as a speaker verification model, this is a case study of speech commands control of mobile robots. The design of the automatic speech recognition system consisted of 3 steps: The first we analyzed the signal processing of the continuous speech commands and compared the accuracy of the speech recognition with a time frame adjustment and the overlapped period of signal filtered with the window function, The second we proceed to find the feature extraction of speech commands using format frequency techniques and configured the feature extraction with format frequencies of F1, F2, and F3,The last step was to design the recognition using Support Vector Machine technique to check the accuracy of an automatic speech recognition. These is support vector machine classification algorithm provides a comparison of the filtered function window and compares the accuracy of the time frame scaled and the overlapped time of the filtered, which gives different values of precision. From the experiment, the researcher found that are applied a Hanging function the test results of the test result of the "forward" speech commands has an accuracy of 81.92% but kind of Gaussian function the test results of the "backward" speech commands has an accuracy of 83.69%, the "turn left" speech commands had an accuracy of 82.81%, the "turn right" speech commands had an accuracy of 85.56% and the "Stop first" speech commands has an accuracy of 86.78% and speech recognition by continuous speech commands recognition with Thai language was applied an every function the test results of the overall performance of the speech commands has an accuracy of 83.88%.

PL

Artykuł przedstawia zastosowanie automatycznego rozpoznawania mowy poprzez ciągłe rozpoznawanie poleceń głosowych z językiem tajskim jako modelem weryfikacji mówiącego, jest to studium przypadku sterowania poleceniami głosowymi robotów mobilnych. Projekt systemu automatycznego rozpoznawania mowy składał się z 3 etapów: W pierwszym przeanalizowano przetwarzanie sygnału ciągłych poleceń głosowych i porównano dokładność rozpoznawania mowy z dopasowaniem przedziału czasowego i nakładającym się okresem sygnału filtrowanego funkcją okna, Następnie przystępujemy do znalezienia ekstrakcji funkcji poleceń głosowych przy użyciu technik formatowania częstotliwości i skonfigurowania ekstrakcji cech z częstotliwościami formatu F1, F2 i F3. Ostatnim krokiem było zaprojektowanie rozpoznawania przy użyciu techniki maszyny wektorów nośnych w celu sprawdzenia dokładności automatyczne rozpoznawanie mowy. Jest to algorytm klasyfikacji maszyny wektorów nośnych, który zapewnia porównanie przefiltrowanego okna funkcji i porównuje dokładność skalowanych ram czasowych oraz nakładających się czasów filtrowanych, co daje różne wartości precyzji. Na podstawie eksperymentu badacz odkrył, że po zastosowaniu funkcji wiszącej wyniki testu wyników poleceń głosowych „do przodu” mają dokładność 81,92%, ale rodzaj funkcji Gaussa wyniki testu poleceń głosowych „wstecz” mają dokładność 81,92% dokładność 83,69%, polecenia głosowe „skręć w lewo” miały dokładność 82,81%, polecenia głosowe „skręć w prawo” miały dokładność 85,56%, a polecenia głosowe „Najpierw zatrzymaj” mają dokładność 86,78%, a rozpoznawanie mowy przez zastosowano ciągłe rozpoznawanie poleceń głosowych w języku tajskim, a wyniki testu ogólnej wydajności poleceń głosowych mają dokładność 83,88%.

15

Advanced extreme learning machine for an hour PV forecast using general weather data

Farid Imam Wahyudi, Priyad Ardyono, Purnomo Mauridhi Hery, Abdillah Muhammad, Tjahjono Anang, Yorino Naoto

Przegląd Elektrotechniczny

|

2023

|

R. 99, nr 1

134--139

EN

In recent years, Indonesia has placed great attention on the use of renewable energy resources as a way to decrease gas emission. Located at the equator, Indonesia has many advantages in renewable energy resources, especially photovoltaic (PV). Photovoltaic offers a big opportunity to contribute to the power grid, yet it also comes with its challenges. The use of PV involves a major uncertainty as the inputs of PV are weather conditions that are constantly changing. With Indonesia planning to penetrate the PV farm into the power grid, it is necessary to be able to generate an accurate forecast to assist the power grid control operator. Many algorithms are applied to obtain a precise and accurate PV power generation. One of the algorithms generally used by researchers is the conventional back propagation neural network. It is one of the most commonly applied algorithms, yet it also has a complex setting and numerous parameters. To help overcome this issue, extreme learning machine (ELM) is applied alongside with backpropagation neural network (BPNN), resulting in a more promising result. However, the random value for ELM parameters has become another problem of its own. This paper discusses an advanced ELM to obtain a better PV forecast result. The combination of PV input, ambient temperature, global tilted irradiation (GTI), wind direction, wind velocity and humidity are applied on the kernel extreme learning machine (K-ELM). We found that K-ELM proposes a better performance compared to ELM in facing a nonlinear data, along with better learning capability, mapping ability, and an improved efficiency. We also developed the input data using BPNN, ELM and support vector machine (SVM) to compare training, testing and calculation time

PL

W ostatnich latach Indonezja przywiązywała dużą wagę do wykorzystania odnawialnych źródeł energii jako sposobu na zmniejszenie emisji gazów. Położona na równiku Indonezja ma wiele zalet w zakresie odnawialnych źródeł energii, zwłaszcza fotowoltaiki (PV). Fotowoltaika daje duże możliwości wniesienia wkładu w sieć energetyczną, ale wiąże się również z wyzwaniami. Korzystanie z PV wiąże się z dużą niepewnością, ponieważ wejścia PV to stale zmieniające się warunki pogodowe. Ponieważ Indonezja planuje penetrację farmy fotowoltaicznej do sieci energetycznej, konieczne jest wygenerowanie dokładnej prognozy, aby pomóc operatorowi kontroli sieci energetycznej. W celu uzyskania precyzyjnego i dokładnego wytwarzania energii PV stosuje się wiele algorytmów. Jednym z algorytmów powszechnie stosowanych przez badaczy jest konwencjonalna sieć neuronowa wstecznej propagacji. Jest to jeden z najpowszechniej stosowanych algorytmów, ale ma też złożoną nastawę i liczne parametry. Aby rozwiązać ten problem, zastosowano ekstremalną maszynę uczącą (ELM) wraz z siecią neuronową z propagacją wsteczną (BPNN), co daje bardziej obiecujący wynik. Jednak losowa wartość parametrów ELM stała się kolejnym problemem. W niniejszym artykule omówiono zaawansowane ELM w celu uzyskania lepszego wyniku prognozy PV. Kombinacja sygnału wejściowego PV, temperatury otoczenia, napromieniowania globalnego odchylenia (GTI), kierunku wiatru, prędkości wiatru i wilgotności jest stosowana na maszynie ekstremalnego uczenia jądra (K-ELM). Odkryliśmy, że K-ELM proponuje lepszą wydajność w porównaniu do ELM w obliczu danych nieliniowych, a także lepszą zdolność uczenia się, zdolność mapowania i lepszą wydajność. Opracowaliśmy również dane wejściowe za pomocą BPNN, ELM i maszyny wektorów nośnych (SVM) w celu porównania czasu szkolenia, testowania i obliczeń.

16

A novel hybrid framework to model the relationship of daily river discharge with meteorological variables

Shabbir Maha, Chand Sohail, Iqbal Farhat

Meteorology Hydrology and Water Management. Research and Operational Applications

|

2023

|

Vol. 11, Iss. 2

70--94

EN

River discharge is affected by many factors, such as water level, rainfall, and precipitation. This study proposes a new hybrid framework named LAES (LASSO-ANN-EMD-SVM) to model the relationship of daily river discharge with meteorological variables. This hybrid framework is a composite of the least absolute shrinkage and selection operator (LASSO), an artificial neural network (ANN), and an error correction method. In the first stage, LASSO identifies meteorological variables that have a significant influence on the generation of river discharge. Next, the ANN model is used to predict river discharge using meteorological variables selected by LASSO, and the error series is determined. The error series is decomposed into intrinsic mode functions and residuals using empirical mode decomposition (EMD). The EMD components are modeled using the support vector machine (SVM) model, and the error predictions are aggregated. In the last stage, the LASSO-ANN predictions and the predicted error series are aggregated as the final discharge prediction. The proposed hybrid framework is illustrated on the Kabul River of Pakistan. The performance of the proposed hybrid framework is compared with six models using various performance measures and the Diebold-Mariano test. These models include multiple linear regression (MLR), SVM, ANN, LASSO-MLR, LASSO-SVM, and LASSO-ANN models. The findings reveal that the proposed hybrid model outperforms all other models considered in the study. In the testing phase, the root mean square error (RMSE), mean absolute percentage error (MAPE), and mean absolute error (MAE) of the proposed LAES hybrid model are 337.143 m3/s, 32.354%, and 218.353 m3/s which are smaller than all other models compared in the study. Our proposed hybrid system is an efficient model for river discharge prediction that will be helpful in water management and protection against floods. Long-term prediction can help to identify the major effects of climate change and to make evidence-based environmental policies.

17

Sentiment analysis of user preference for old vs new fintech technology using SVM and NB algorithms

Nurdin Tubagus Asep, Alexandri Mohammad Benny, Sumadinata Widya, Arifianti Ria

Management Systems in Production Engineering

|

2023

|

nr 4 (31)

373--380

EN

The aim of this study is to use sentiment analysis to compare the efficiency of old and new fintech technologies by collecting data from various sources and analyzing it using the SVM and NB algorithms. The study seeks to identify opinions or feelings from text in order to provide a clear picture of public opinion and the direction of the debate regarding old and new fintech technologies. The results of the study show that the SVM algorithm has an average accuracy of 87.32% and the NB algorithm has an average accuracy of 81.56% in testing the sample data in a comparison of old and new fintech technology on the internet. The study tested data in a comparison of two specific arguments, namely the debate about which technology is more efficient in old and new fintech on the internet. Despite many unresolved arguments, the study successfully proved that new fintech is more preferred than old fintech, with 71% positive sentiment directed towards new fintech. However, the dataset also found that 62% negative sentiment is directed towards new fintech, indicating that although new fintech is more preferred, there are still some issues that need to be addressed. One reason for negative sentiment towards new fintech may be the continued concerns about security and privacy of user data. Furthermore, other factors that may cause negative sentiment towards new fintech include a lack of understanding about how the technology works.

18

Enhancing Multiclass Pneumonia Classification with Machine Learning and Textural Features

Godbin A. Beena, Jasmine S. Graceline

Machine Graphics and Vision

|

2023

|

Vol. 32, No. 3/4

83--106

EN

The highly infectious and mutating COVID-19, known as the novel coronavirus, poses a substantial threat to both human health and the global economy. Detecting COVID-19 early presents a challenge due to its resemblance to pneumonia. However, distinguishing between the two is critical for saving lives. Chest X-rays, empowered by machine learning classifiers and ensembles, prove effective in identifying multiclass pneumonia in the lungs, leveraging textural characteristics such as GLCM and GLRLM. These textural features are instilled into the classifiers and ensembles within the domain of machine learning. This article explores the multiclass categorization of X-ray images across four categories: COVID-19-impacted, bacterial pneumonia-affected, viral pneumonia-affected, and normal lungs. The classification employs Random Forest, Support Vector Machine, K-Nearest Neighbor, LGBM, and XGBoost. Random Forest and LGBM achieve an impressive accuracy of 92.4% in identifying GLCM features. The network’s performance is evaluated based on accuracy, precision, sensitivity and F1-score.

19

Comparison of Machine Learning Algorithms on Classification of Covid-19 Cough Sounds Using MFCC Extraction

Faisal Mohammad Reza, Hidayat Muhammad Thoriq, Kartini Dwi, Indriani Fatma, Budiman Irwan, Saragih Triando Hamonangan

Journal of Computer Sciences Institute

|

2023

|

Vol. 29

399--404

EN

Early detection for COVID-19 has now been widely developed. One of the methods used is cough audio detection. This research aims to classify cough audio. Audio feature extraction is performed using MFCC to obtain numerical features. Feature classification is done using SVM, Random Forest, and Naive Bayes methods. Evaluation is done to find the best classification method. The evaluation results in this study show that SVM Kernel RBF produces the best evaluation value with an AUC value of 0.657715.

PL

Wczesne wykrywanie COVID-19 zostało obecnie szeroko opracowane. Jedną z zastosowanych method jest wykrywanie dźwięku kaszlu. Badania te mają na celu klasyfikację dźwięku kaszlu. Ekstrakcję próbek audio wykonano przy użyciu Mel Frequency Cepstral Coefficients (MFCC) w celu uzyskania cech numerycznych. Klasyfikacja cech odbywa się przy użyciu metod SVM, Random Forest i Naive Bayes. Wyniki oceny w tym badaniu pokazują, że SVM Kernel RBF daje najlepszą wartość oceny z wartością AUC wynoszącą 0.657715.

20

An Efficient Hybrid Classifier Model for Customer Churn Prediction

Anitha M. A., Sherly K. K.

International Journal of Electronics and Telecommunications

|

2023

|

Vol. 69, No. 1

11--18

EN

Customer churn prediction is used to retain customers at the highest risk of churn by proactively engaging with them. Many machine learning-based data mining approaches have been previously used to predict client churn. Although, single model classifiers increase the scattering of prediction with a low model performance which degrades reliability of the model. Hence, Bag of learners based Classification is used in which learners with high performance are selected to estimate wrongly and correctly classified instances thereby increasing the robustness of model performance. Furthermore, loss of interpretability in the model during prediction leads to insufficient prediction accuracy. Hence, an Associative classifier with Apriori Algorithm is introduced as a booster that integrates classification and association rule mining to build a strong classification model in which frequent items are obtained using Apriori Algorithm. Also, accurate prediction is provided by testing wrongly classified instances from the bagging phase using generated rules in an associative classifier. The proposed models are then simulated in Python platform and the results achieved high accuracy, ROC score, precision, specificity, F-measure, and recall.