Initial Value Selection Research Articles

From machine learning to deep learning: experimental comparison of machine learning and deep learning for skin cancer image segmentation

Skin lesion analysis is a tedious and challenging task, thus, in this research the suitability of employing machine learning or deep learning approaches for automatic lesion segmentation on dermoscopic skin cancer images is determined. The segmented region can assist clinical experts in understanding the complex lesion structure and internal pattern to find the correct skin cancer type for its early diagnosis and prevention. In this study, I present two methodologies for performing lesion segmentation: machine learning-based optimized K-means with Firefly Algorithm (FA) and Convolutional Neural Network (CNN). In the first model, the FA is hybridized with K-means clustering based on the novel average intensity fitness function to optimize the segmentation map. It is observed in the experimental results that the K-means algorithm may lead to poor results due to the wrong selection of initial centroid value, thus FA is hybridized into it to improve the performance. The second model is an enhanced encoder-decoder-based CNN framework implemented in an end-to-end fashion. These two models are compared to understand whether machine learning or deep learning is suitable to perform medical image segmentation based on a few performance metrics such as accuracy, Intersection over Union (IoU), and DICE index. These methods are evaluated and compared on two benchmark datasets provided by the International Skin Imaging Collaboration (ISIC) named ISIC 2016 [1] and ISIC 2017 [2]. Experimental results showed that the CNN model outperformed the machine learning model with an accuracy difference of 7.98% on ISIC 2016, and 7.32% on ISIC 2017. I concluded from the experimental findings that the deep learning model is more accurate and efficient in segmenting the lesion area as compared to the machine learning model. Thus, findings from this experimental work will be considered for the design of an automatic classification system by incorporating a deep learning-based segmentation approach as a pre-processing step.

An Inverse Solution Algorithm for Industrial Robot

The inverse solution process of industrial robot is complex. Because the traditional matrix algorithm involves a lot of matrix operation, its process is complex. Numerical algorithms usually do not yield all the solutions, generally, the selection of initial value and the search algorithm have great influence on the convergence and accuracy. The geometric algorithms is simple and intuitionistic with simple expression and small computation. In this paper, a hybrid algorithm combining the geometric method and the inverse transformation matrix method is proposed for puma 560 industrial robots. The geometric algorithms is used to solve the first three joints and the matrix algorithm is used to obtain the final three joint angles. This paper introduces the inverse solution algorithm for this kind of robot in detail, and verifies the effectiveness of the proposed algorithm through matlab simulation.

A Point Cloud Fusion Method for Space Target 3D Laser Point Cloud and Visible Light Image Reconstruction Point Cloud

In this paper, a high-precision spatial target lidar for 3D reconstruction of point cloud and visible light image 3D reconstruction point cloud fusion method is proposed. This fusion method uses the solved 3D reconstruction to invert the camera pose and 3D point cloud model centroid position optimization initial value selection. The registration accuracy and registration efficiency of the ICP algorithm are improved. At the same time, according to the characteristics of the two sets of point clouds, the Euclidean distance threshold is used to delete the noise points of the 3D point cloud edge, and the high-precision 3D reconstruction point with the scale information fusion is obtained. cloud. The simulation experiment of the spatial target simulation model shows that the fusion method can effectively improve the point cloud density, fill the reconstruction vulnerability, and improve the point cloud accuracy of the spatial target 3D reconstruction.

Quantitative analysis of internal polarization dynamics for polymer electrolyte membrane fuel cell by distribution of relaxation times of impedance

Investigating and interpreting each internal polarization dynamics that occurs in the polymer electrolyte membrane fuel cell is significant. Traditional equivalent circuit model fitting by nonlinear least-squares relies on prior model assumptions and initial value selection of components. In this paper, the distribution of relaxation times methodology with powerful separating ability is applied to reveal a more precise analysis of polarization processes. First, the electrochemical impedance spectroscopy under a broad of operating conditions is carried out. Four polarization dynamics related to oxygen transfer, charge transfer of the oxygen reduction, proton transfer inside cathode ionomer, and interface contact process between catalyst layer and membrane (perhaps, including anode oxidation reaction) are effectively extracted. Then, a fourth-order equivalent circuit model established via distribution of relaxation times results is introduced to quantify the loss of each polarization process. Based on this, for the first time, the sensitivity of each polarization loss against operating conditions is analyzed by the multiple stepwise regression analysis, and its application on vehicular fuel cell system control is discussed. Afterward, the distribution of relaxation times is also first to explore the loss and variation trend of each polarization process under flooding, membrane drying, and air starvation fault, where each failure type contains at least eight test sequences. These efforts represent a comprehensive and systematic guideline for fuel cells using distribution of relaxation times, which can also guide the study of degradation mechanisms, optimization design of materials, and even other electrochemical energy sources.

A Novel Quaternion Kernel LMS Algorithm With Variable Kernel Width

The quaternion Gaussian kernel is usually used when solving quaternion nonlinear problems. However, how to choose a proper value of kernel width is still an important issue. In most previous studies, the kernel width was set manually or estimated in advance by using Silvermans rule based on the sample distribution, which can easily degrade the performance of algorithms. In this brief, the variable kernel width quaternion kernel least mean squares algorithm (VKW-QKLMS) aims to develop an online technique for optimizing the kernel width of the quaternion kernel LMS (QKLMS) algorithm, in which the filter weight and the kernel width are alternately updated by using stochastic gradient algorithm. Simulation results show that the performance of the VKW-QKLMS algorithm does not depend on the selection of initial value of the kernel width. The VKW-QKLMS algorithm can usually achieve better performances than other competing algorithms. Only when the kernel width in the QKLMS algorithm is empirically selected as a proper value, the QKLMS algorithm can achieve much the same performance as that of the VKW-QKLMS algorithm.

Static bifurcation and vibrational resonance in an asymmetric fractional-order delay Duffing system

Static bifurcation and vibrational resonance in a fractional-order delay Duffing system with asymmetric potential are studied. By utilizing the fast and slow variable separation approach, the static bifurcation is discussed based on the equivalent system. It is found when fractional order and time delay change, the static bifurcation of fractional-order system with asymmetric potential is different from that of system with symmetric potential. In addition, changes in asymmetric parameter can also induce saddle-node bifurcation. Furthermore, according to the approximate analytical expression of response amplitude, the effects of high frequency amplitude, fractional order and asymmetric parameter on vibrational resonance are investigated. The results show that the number of resonance peaks of asymmetric potential system is related to the selection of initial value and the change of asymmetric parameter can also induce vibrational resonance. In the meantime, the good agreement between theoretical prediction and numerical simulation confirms the validity of theoretical analysis.

An improved multi-node Newton iteration search method based on PSO

When using the search algorithm to locate the vibrating target, Newton iteration method has strict requirement on the selection of initial value. To solve this problem, an improved Newton iterative search algorithm for vibration source location is proposed based on the analysis of Newton iterative search algorithm for multi-node co-location. The algorithm uses the estimation characteristics of POS algorithm to get the initial position of vibration source. Then the position of vibration target is determined by Newton iterative search method. The experimental results show that this method can effectively solve the problem of initial value selection of Newton iteration method and improve the localization accuracy of vibration source.

Fetal Electrocardiography Extraction Based on Improved Fast Independent Components Analysis Algorithm.

Aiming at the problem that the fast independent components analysis (ICA) algorithm is sensitive to the initial value and easy to fall into local optimization, an improved fast ICA method is proposed for the extraction of fetal echocardiography (FECG) by combining the Simpson-Newton iterative and chaotic optimization algorithm to replace the traditional Newton iterative method. First, the Simpson formula is used to modify the traditional Newton method and a Simpson-Newton iterative algorithm is constructed. It shows that the Simpson-Newton iterative algorithm can significantly reduce the sensitivity of initial value selection, and has faster convergence speed. Then, combined with the Simpson-Newton method, the chaos optimization algorithm can obtain the approximate global optimal solution, which solves the problem that the traditional Newton iterative method tends to fall into local optimal and improves the separation performance of the fast ICA algorithm. Finally, based on chaos optimization, the proposed Simpson-Newton iterative fast ICA algorithm is applied to the extraction of FECG signals, and the extraction effect is evaluated by visual waveform and quantitative indicators. Furthermore, the algorithm is verified by different clinical signals. The experimental results show that the improved fast ICA algorithm can extract clear fetal heart signals, and there are almost no mixed maternal ECG signals in the extracted FECG signals. The extraction effect of the proposed method is thus optimal than that of the traditional fast ICA method.

Hybrid Location Algorithm Based on Cuckoo and Statistical Manifold

To improve the location accuracy, a hybrid location algorithm based on cuckoo and statistical manifold method is proposed. It combines the cuckoo algorithm's strong global optimization ability and the statistical manifold’s accurate positioning ability fully. The simulation results show that the hybrid location algorithm has higher accuracy and reduces the influence of initial value selection on location accuracy.

Research on Inverse Kinematics Algorithm of 6-DOF Industrial Robot Based on RBF-PID

To solve the question of multi value and precision in the process of designing the inverse kinematics controller of 6-DOF Industrial robot, a radial basis function-proportional integral differentiation(RBF-PID) adaptive controller method has been proposed to train the inverse kinematics model of industrial robot. Compared with the traditional RBF neural network, the advantage of RBF-PID adaptive controller was that it had high search efficiency and was not easy to fall into local optimal solution. The simulation results showed that this method could not only overcome the influence of the initial value selection of the traditional RBF neural network learning algorithm on the network performance, but also had a good solution accuracy for the inverse kinematics of the 6-DOF Industrial robot. Compared with the traditional RBF neural network algorithm, the RBF-PID adaptive controller method had improved the performance of overshoot, adjustment time and anti-interference, and solved the problem of solving the multivalued inverse kinematics algorithm.

Bearing-Only Target Tracking Based on Piecewise Backward Smoothing Kalman Filter

In order to solve the problem that the Unscented Kalman Filter algorithm is sensitive to the initial value selection and the noise is easy to expand the error in the recursive process, a piecewise backward smoothing Kalman filter method is proposed in this paper. To reduce the influence of noise, the observation data is optimized in sections, and the forward and reverse use of historical data is increased. The simulation results under different initial distance error, different observation angle variance and different initial velocity error show that the proposed method is effective in underwater bearings only target tracking, and the effect of Unscented Kalman filter is obviously improved, the estimation error is greatly reduced, and the estimation accuracy and stability are improved.

Hybrid PIPSO-SQP Algorithm for Real Power Loss Minimization in Radial Distribution Systems with Optimal Placement of Distributed Generation

This paper proposes the hybrid sequential quadratic programming (SQP) technique based on active set method for identifying the optimal placement and rating of distribution generation (DG) incorporated in radial distribution systems (RDS) for minimizing the real power loss satisfying power balance equations and voltage limits. SQP runs quadratic programming sequentially as a sub-program to obtain the best solution by using an active set method. In this paper, the best optimal solution is selected with less computation time by combining the benefits of both classical and meta-heuristic methods. SQP is a classical method that is more sensitive to initial value selection and the evolutionary methods give approximate solution. Hence, the initial values for the SQP technique were obtained from the meta–heuristic method of Parameter Improved Particle Swarm Optimization (PIPSO) algorithm. The proposed hybrid PIPSO–SQP method was implemented in IEEE 33-bus RDS, IEEE 69-bus RDS, and IEEE 118-bus RDS under different loading conditions. The results show that the proposed method has efficient reduction in real power loss minimization through the enhancement of the bus voltage profile.

Enhanced Computation Performance of Photovoltaic Models for Power Hardware-in-the-Loop Simulation

For power hardware-in-the-loop (PHIL) simulation, a real-time simulator has to complete the target model calculations in a real-time manner without overrun errors. However, a photovoltaic (PV) simulation model contains a nonlinear equation that demands a numerical method with long computation time. Therefore, the complexity and operating condition of the target PV model are limited to reduce its computation burden. Besides, it sacrifices the model accuracy and limits the simulation scenario of the PHIL simulation. In this article, the PV simulation model employing an effective initial value selection method is proposed to enhance the real-time simulation performance for the PV PHIL simulation. The proposed initial value selection method can reduce the number of iterations of the numerical method so that it can drastically reduce the computation time of the real-time simulation. Therefore, the PHIL simulation model can increase it complexity with a fixed time step. Moreover, the PV model can be scaled up with various operating conditions, which can increase the PHIL simulation accuracy. The accuracy and the performance of the proposed PV model are evaluated by the PHIL simulation results. The maximum number of the PV models for the target PHIL simulation is also discussed.

A Revised Point-to-Point Calibration Approach with Adaptive Errors Correction to Weaken Initial Sensitivity of Cuff-Less Blood Pressure Estimation.

Initial calibration is a great challenge for cuff-less blood pressure (BP) measurement. The traditional one point-to-point (oPTP) calibration procedure only uses one sample/point to obtain unknown parameters of a specific model in a calm state. In fact, parameters such as pulse transit time (PTT) and BP still have slight fluctuations at rest for each subject. The conventional oPTP method had a strong sensitivity in the selection of initial value. Yet, the initial sensitivity of calibration has not been reported and investigated in cuff-less BP motoring. In this study, a mean point-to-point (mPTP) paring calibration method through averaging and balancing calm or peaceful states was proposed for the first time. Thus, based on mPTP, a factor point-to-point (fPTP) paring calibration method through introducing the penalty factor was further proposed to improve and optimize the performance of BP estimation. Using the oPTP, mPTP, and fPTP methods, a total of more than 100,000 heartbeat samples from 21 healthy subjects were tested and validated in the PTT-based BP monitoring technologies. The results showed that the mPTP and fPTP methods significantly improved the performance of estimating BP compared to the conventional oPTP method. Moreover, the mPTP and fPTP methods could be widely popularized and applied, especially the fPTP method, on estimating cuff-less diastolic blood pressure (DBP). To this extent, the fPTP method weakens the initial calibration sensitivity of cuff-less BP estimation and fills in the ambiguity for individualized calibration procedure.

Novel Fractional order Integrators Using Traditional Rules Interpolation

This work proposes a new IIR integer order digital, so a new IIR fractional order integrators. The followed method includes two stages. First, the integer order integrator is obtained by interpolating well-known integration rulesnamely, Euler, bilinear and Simpsonas a weighted sum. Second the initial value theorem for the selection of initial value of the impulse response is used. Lastlythe Steiglitz-Mc-bride signal modeling technique is applied to find the parameters of rational approximation models. Numerical examples are presented to illustrate the performance of the proposed integrator. It was found that the Euler-bilinear-Simpson integrator yields high accuracy than the existing integrators.

Using genetic algorithms for estimating Weibull parameters with application to wind speed

Renewable energy has become a prominent subject for researchers since fossil fuel reserves have been decreasing and are not promising to meet the energy demand of the future. Wind takes an important place in renewable energy resources and there is extensive research on wind speed modeling. Herein, one of the most commonly used distributions for wind speed modeling is the Weibull distribution with its simplicity and flexibility. Maximum likelihood (ML) method is the most frequently used technique in Weibull parameter estimation. Iterative techniques such as Newton-Raphson (NR) use random initial values to obtain the ML estimators of the parameters of the Weibull distribution. Therefore, the success of the iterative techniques highly depends on the initial value selection. In order to deliver a solution to the initial value problem, genetic algorithm (GA) is considered to obtain the estimators of the model parameters. The ML estimators obtained using the GA and NR techniques are compared with the method of moments (MoM) estimators via Monte Carlo simulation and wind speed applications. The results show that the ML estimators obtained using GA present superiority over MoM and the ML estimators obtained using NR.

Quality over quantity: on workflow and model space exploration of 3D inversion of MT data

3D inversions of magnetotelluric data are now almost standard, with computational power now allowing an inversion to be performed in a matter of days (or hours) rather than weeks. However, when compared to 2D inversions, these are still very computationally demanding. As a result, 3D inversions are generally not subjected to as rigorous testing as a 1D or 2D inversion would be, which has implications when these models are used for geological interpretation. In this study, we explore the parameter space for inversion of continent-scale datasets. The generalisations made regarding the effects of each parameter should also be scalable to smaller surveys and will enable MT practitioners to optimise their results. We have performed testing on a subset of the South Australian component of the eventual Australia-wide AusLAMP (Australian Lithospheric Architecture Magnetotelluric Project). The subset was inverted with different parameters, model setup and data subsets. Specifically, results from testing of the model covariance, the resistivity of the prior model, the inclusion of 'known' information into the prior model, the model cell size, the data components inverted for and the damping parameter lambda were all investigated. In our testing of the 3D inversion software, ModEM3DMT, we found that the resistivity of the starting/prior model had significant effect on the final model. Careful selection of initial lambda value can aid in reducing computational time whilst having a negligible effect on the resultant model, whilst large covariance values and model cell sizes enhanced conductive features at depth.

The Study of an Improved Text Clustering Algorithm for Self-Organizing Maps

The traditional SOM algorithm need to determine the number of clustering categories in advance, which is very subjective. In this paper, an improved k-means initial value selection algorithm is proposed to calculate the number of clustering categories, which is applied to SOM network model. In this algorithm, the Latent Semantic Indexing is applied in the pre-processing stage of clustering, and the improved SOM algorithm is applied in the text clustering stage. Namely, the number of clustering categories obtained by the improved k-means initial value selection algorithm is taken as the number of neurons in the output layer of SOM network.

A novel FastICA algorithm based on improved secant method for Intelligent drive

Blind Source Separation(BSS) is one of the research hotspots in the field of signal processing. In order to improve the accuracy of speech recognition in driving environment, the driver’s speech signal must be enhanced to improve its signal to noise ratio(SNR). Independent component analysis (ICA) algorithm is the most classical and efficient blind statistical signal processing technique. Compared with other improved ICA algorithms, fixed-point algorithm (FastICA) is well known for its fast convergence speed and good robustness. However, the convergence of FastICA algorithm is comparatively susceptible to the initial value selection of the original demixing matrix and the calculation of the iterative process is relatively large. In this paper, the gradient descent method is used to reduce the effect of initial value. What’s more, the improved secant method is proposed to speed up the convergence rate and reduce the amount of computation. As the results of mixed speech separation experiment turn out, the improved algorithm is of better performance relative to the standard FastICA algorithm. Experimental results show that the proposed algorithm improves the speech quality of the target driver. It is suitable for speech separation in driving environment with low SNR.

Influence of EEDI (Energy Efficiency Design Index) on Ship–Engine–Propeller Matching

With the increasingly strict international GHG (greenhouse gas) emission regulations, higher requirements are placed on the propulsion system design of conventional ships. Playing an important role in ship design, construction and operation, ship–engine–propeller matching dominantly covers the CO2 emission of the entire ship. In this paper, firstly, a ship propulsion system matching platform based on the ship–engine–propeller matching principle and its application on WinGD 5X52 marine diesel engine have been investigated. Meeting the energy efficiency design index (EEDI) regulation used to calculate the ship CO2 emission is essential and ship–engine–propeller matching has to be carried out with EEDI into consideration. Consequently, a procedure is developed combining the system matching theory and EEDI calculation, which can provide the matching results as well as the corresponding EEDI value to study the relationship between EEDI and ship–engine–propeller matching. Furthermore, a comprehensive analysis is performed to obtain the relationship of EEDI and system matching parameters, such as ship speed, effective power and propeller diameter, reflecting the trend and extent of EEDI when changing these three parameters. The results of system matching parameters satisfying different EEDI phases indicate the initial value selection in matching process to provide reference for the design of ship, engine and propeller under the EEDI regulations.