Wyniki wyszukiwania - BazTech

1

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.

2

Performance Evaluation of the Threshold Call Admission Policy in Multi-rate Loss Systems

Moscholios Ioannis D., Chousainov Iskanter-Alexandros, Panagoulias Panagiotis I., Sarigiannidis Panagiotis G., Logothetis Michael D.

Journal of Telecommunications and Information Technology

|

2020

|

nr 2

51--60

EN

In this paper we consider a link, characterized by speciﬁc capacity, that services multi-rate random or quasirandom traﬃc. Random traﬃc is generated by an inﬁnite number of traﬃc sources, while quasi-random traﬃc is generated by a ﬁnite population of traﬃc sources. The link is modeled as a multi-rate loss system. Handover and new calls are distinguished. New calls compete for the available bandwidth under a threshold call admission policy. In that policy, a new call of a particular service-class is not allowed to enter the system if the in-service handover and new calls of the same service-class plus the new call, exceed a predeﬁned threshold (which can be diﬀerent for each service-class). On the other hand, handover calls compete for the available bandwidth based on the complete sharing policy. We show that the steady state probabilities in the proposed models have a product form solution (PFS). The PFS leads to a convolution algorithm for accurate calculation of congestion probabilities and link utilization.

3

Proposed Hypothesis on replacing Non-Deterministic Approach to Abstraction due to the exhibition of faulty speculations

Indukuri Shreya, Konda Anil Kumar Reddy, Sharma Mellacheruvu Janakirama

Annals of Computer Science and Information Systems

|

2020

|

Vol. 24

69--72

EN

Non-Determinism is the word that is used to refer to an event or a series of the same that is said to be unpredictable but is assumed to be belonging to a domain of possibilities. From the inception, non-deterministic approaches have played a huge role in statistics, machine design, the definition of Random, and a few other domains to mention. Non-Deterministic processes are defined as the paths that lead to no fixed state but a set of possible states instead. One of the crucial reasons would be delving into the human understanding of processes and the ability to justify happenings around. A behavior that has accumulated throughout the evolution of mankind is to explain boundaries based on what is being perceived. The argument of Human Understanding of processes around can be of huge prominence as they play a vital role in explaining and demystifying a lot of misconceptions. One such aspect that is a resultant of this phenomenon would be a nondeterministic approach. The key focus of this paper will be on emphasizing how humans are limited to perceiving reality, how several misconceptions have accumulated around misstated definitions by the virtue of the same, have a closer look at Randomness, a sheer epitome of the same, and also propose a few potential applications that are affected by the hitherto stated problem.

4

On wave propagation in a random micropolar generalized thermoelastic medium

Mitra M., Bhattacharyya R. K.

Archives of Thermodynamics

|

2017

|

Vol. 38, no. 2

21--60

EN

This paper endeavours to study aspects of wave propagation in a random generalized-thermal micropolar elastic medium. The smooth perturbation technique conformable to stochastic differential equations has been employed. Six different types of waves propagate in the random medium. The dispersion equations have been derived. The effects due to random variations of micropolar elastic and generalized thermal parameters have been computed. Randomness causes change of phase speed and attenuation of waves. Attenuation coefficients for high frequency waves have been computed. Second moment properties have been briefly discussed with application to wave propagation in the random micropolar elastic medium. Integrals involving correlation functions have been transformed to radial forms. A special type of generalized thermo-mechanical auto-correlation functions has been used to approximately compute effects of random variations of parameters. Uncoupled problem has been briefly outlined.

5

Calibration of characteristic values of soil properties using the random finite element method

Zaskórski Ł., Puła W.

Archives of Civil and Mechanical Engineering

|

2016

|

Vol. 16, no. 1

112--124

EN

In this paper the reliability assessment of the shallow strip footing was conducted using a reliability index β. Therefore some approaches of evaluation of characteristic values of soil properties were compared in order to check what reliability index β can be achieved by applying each of them. For this purpose, design values of the bearing capacity based on these approaches were referred to design values of the bearing capacity estimated by the random finite element method. Design values of the bearing capacity were estimated for various widths and depths of foundation in conjunction with design approaches defined in Eurocode 7. The cohesive soil was considered – clay from the area of Wrocław. The characteristic values of shear strength parameters were evaluated basing on the effective values of soil parameters.

6

Stochastic modelling and analysis of IMU sensor errors

Zhao Y., Horemuz M., Sjöberg L. E.

Archiwum Fotogrametrii, Kartografii i Teledetekcji

|

2011

|

Vol. 22

437--449

EN

The performance of a GPS/INS integration system is greatly determined by the ability of stand-alone INS system to determine position and attitude within GPS outage. The positional and attitude precision degrades rapidly during GPS outage due to INS sensor errors. With advantages of low price and volume, the Micro Electrical Mechanical Sensors (MEMS) have been wildly used in GPS/INS integration. Moreover, standalone MEMS can keep a reasonable positional precision only a few seconds due to systematic and random sensor errors. General stochastic error sources existing in inertial sensors can be modelled as (IEEE STD 647, 2006) Quantization Noise, Random Walk, Bias Instability, Rate Random Walk and Rate Ramp. Here we apply different methods to analyze the stochastic sensor errors, i.e. autoregressive modelling, Gauss-Markov process, Power Spectral Density and Allan Variance. Then the tests on a MEMS based inertial measurement unit were carried out with these methods. The results show that different methods give similar estimates of stochastic error model parameters. These values can be used further in the Kalman filter for better navigation accuracy and in the Doppler frequency estimate for faster acquisition after GPS signal outage.

7

A View on The Ways Design of Reliability Criteria in Structural Mechanics

Kopecký M.

Machine Dynamics Research

|

2010

|

Vol. 34, No. 3

124-132

EN

Contemporary intensive development of technology puts ever-increasing demands on the reliability of products. The increase in the reliability level is emphasised also in transport machines and equipment. This all requires a further improvement of the method of designing and strength checking of a construction. The methods described in this paper are the ways to reach the solution goals with the maximum use of computer technology. A practical example of loading system analysis in presented which demonstrates use the special instrument to measurement of distribution the force and torsion moment in card an-joint for control purposes and uses the special instrument to generally measurement of distribution random loading parameter. In some machinery and equipment, or their elements, the problem of strength reliability is due to the present-day regulations conditioned by a fatigue process and by knowledge of a fatigue curve. A considerable part of dynamically loaded components in mechanical engineering is loaded with time-variable strength. In the case of the means of machinery , competitiveness leads to weight reduction savings and extreme operational situations cause high overloading. It is probable that the most significant cause of this unfortunate situation is the fact that so far there has not been a single universal theoretically and experimentally proved fatigue theory which would consider all the factor that have an influence on the phenomena. Apart from this, there is not a generally accepted methodology of fatigue tests. A demand to guarantee strength reliability is more important predominantly in case when a failure-free operation of a constructional element can influence the safety of human lives or when an eventual failure can bring about considerable economic losses. The variability of fatigue mechanisms in the Wohler curve belongs among the questions so far not clearly specified. It has to be said, that insufficient information about the values of the fatigue curve can lead to errors sometimes higher than 100%. Also, because of economic and others reasons, it is impossible to carry out sufficient amount of fatigue tests on a finished product in order to get reliable information about the fatigue curve. It is further connected with a choice of a suitable commutation of a fatigue failure which can also lead to different results. Random operational loading creates a stochastic process of excitation forces. A successful reproduction of the response of this random loading depends on the technical facilities. With regard to a probability character of the fatigue curve, where its exponents p, q, have a character of a random quantity, a total service life is also a random quantity characterized by a probability density. The probability that a failure might occur before a required servicelife Lp ends is given by a hatched surface. The methods described in this paper are the ways to reach the solution goals by means of a characteristic curve of strength reliability and reduced fatigue curve with the maximum use of computer technology. The application of this methodology shortens knowledge of the time to failure of mobile machine components and contributes to the safety and economy of mechanical systems. The results of its application would be presented to mobile facility elements.

8

A calculus on Lévy exponents and selfdecomposability on Banach spaces

Jurek Z. J.

Probability and Mathematical Statistics

|

2008

|

Vol. 28, Fasc. 2

271--280

EN

In infinite-dimensional Banach spaces there is no complete characterization of the Lévy exponents of infinitely divisible probability measures. Here we propose a calculus on Lévy exponents that is derived from some random integrals. As a consequence we prove that each selfdecomposable measure can by factorized as another selfdecomposable measure and its background driving measure that is s-selfdecomposable. This complements a result from the paper of Iksanov, Jurek and Schreiber in the Annals of Probability (2004).

9

Application of Adomian method on the solution of wave propagation in a linear random non-homogeneous viscoelastic semi-infinite rod

Ray S. S., Bera R. K.

International Journal of Applied Mechanics and Engineering

|

2005

|

Vol. 10, no 1

133-143

EN

The aim of the present problem is to investigate the efficiency of the method of Adomian for the solution of non-linear and complicated differential equation in a random medium. Here the problem is connected with the investigation of the mean and variance of the displacement distribution in a thin linear random non-homogeneous Biot type viscoelastic semi-infinite rod, due to general time-dependent displacement input at the rod. A truncated series solution of the wave problem following the method of Adomian after using the Laplace transform is obtained for small random variations in viscoelastic properties. Three specific cases concerning the probability measure as a function of the continuous type of random variable have been discussed.