Notes on the H-measure of classifier performance

The H-measure is a classifier performance measure which takes into account the context of application without requiring a rigid value of relative misclassification costs to be set. Since its introduction in 2009 it has become widely adopted. This paper answers various queries which users have raised since its introduction, including questions about its interpretation, the choice of a weighting function, whether it is strictly proper, its coherence, and relates the measure to other work.

GEOGRAPHICALLY WEIGHTED NEGATIVE BINOMIAL REGRESSION UNTUK MENANGANI OVERDISPERSI PADA JUMLAH PENDUDUK MISKIN

Poverty is a condition that is often associated with needs, difficulties an deficiencies in various life circumstances. The number of poor people in Indonesia increase in 2020. This research focus on modelling the number of poor people in Indonesia using Geographically Weighted Negative Binomial Regression (GWNBR) method. The number of poor people is count data, so analysis used to model the count data is poisson regression. If there is overdispersion, it can be overcome using negative binomial regression. Meanwhile to see the spatial effect, we can use the Geographically Weighted Negative Binomial Regression method. GWNBR uses a adaptive bisquare kernel for weighting function. GWNBR is better at modelling the number of poor people because it has the smallest AIC value than poisson regression and negative binomial regression. While the GWNBR method obtained 13 groups of province based on significant variables.

Microdosimetric characterization of a clinical proton therapy beam: comparison between simulated lineal energy distributions in spherical water targets and experimental measurements with a silicon detector

Objective Treatment planning based on computer simulations were proposed to account for the increase in the relative biological effectiveness (RBE) of proton radiotherapy beams near to the edges of the irradiated volume. Since silicon detectors could be used to validate the results of these simulations, it is important to explore the limitations of this comparison. Approach Microdosimetric measurements with a MicroPlus Bridge V2 silicon detector (thickness = 10 µm) were performed along the Bragg peak of a clinical proton beam. The lineal energy distributions, the dose mean values, and the RBE calculated with a biological weighting function were compared with simulations with PHITS (microdosimetric target = 1 µm water sphere), and published clonogenic survival in vitro RBE data for the V79 cell line. The effect of the silicon-to-water conversion was also investigated by comparing three different methodologies (conversion based on a single value, novel bin-to-bin conversions based on SRIM and PSTAR). Main results Mainly due to differences in the microdosimetric targets, the experimental dose-mean lineal energy and RBE values at the distal edge were respectively up to 53% and 28% lower than the simulated ones. Furthermore, the methodology chosen for the silicon-to-water conversion was proven to affect the dose mean lineal energy and the RBE10 up to 32% and 11% respectively. The best methodology to compensate for this underestimation was the bin-to-bin silicon-to-water conversion based on PSTAR. Significance This work represents the first comparison between PHITS-simulated lineal energy distributions in water targets and corresponding experimental spectra measured with silicon detectors. Furthermore, the effect of the silicon-to-water conversion on the RBE was explored for the first time. The proposed methodology based on the PSTAR bin-to-bin conversion appears to provide superior results with respect to commonly used single scaling factors and is recommended for future studies.

Real-Time Exercise Mode Identification with an Inertial Measurement Unit for Smart Dumbbells

Exercise is good for health, quality of life, and maintenance of human muscles. Dumbbells are popular indoor exercise equipment with several benefits such as low cost, high flexibility in space and time, easy operation, and suitability for people of all ages. Facilitated by advances in the Internet of Things, smart dumbbells that provide automatic counting and motion monitoring functions have been developed. To perform these tasks, the key process is identification of exercise mode. This study proposes a method to identify essential muscle groups’ (biceps, triceps, and deltoids) exercise modes of a dumbbell using an inertial measurement unit to provide three-axis angular velocities and accelerations. The motion angles were estimated from the axial acceleration and angular velocity. Phase diagrams and time plots of the axial angle, angular velocity, and acceleration were used to extract significant features of each exercise. Machine Learning and weighting functions were developed to combine these features into an identification index value for accurate identification and classification of the exercise modes. An algorithm was developed to verify the exercise mode identification. The results show that the proposed method and weighting function can successfully identify the six exercise modes. The identification algorithm was 99.5% accurate. The exercise mode identification of the dumbbell is confirmed.

A new weighting scheme for arc based circle cone-beam CT reconstruction

In this paper, we present an arc based fan-beam computed tomography (CT) reconstruction algorithm by applying Katsevich’s helical CT image reconstruction formula to 2D fan-beam CT scanning data. Specifically, we propose a new weighting function to deal with the redundant data. Our weighting function ϖ ( x _ , λ ) is an average of two characteristic functions, where each characteristic function indicates whether the projection data of the scanning angle contributes to the intensity of the pixel x _ . In fact, for every pixel x _ , our method uses the projection data of two scanning angle intervals to reconstruct its intensity, where one interval contains the starting angle and another contains the end angle. Each interval corresponds to a characteristic function. By extending the fan-beam algorithm to the circle cone-beam geometry, we also obtain a new circle cone-beam CT reconstruction algorithm. To verify the effectiveness of our method, the simulated experiments are performed for 2D fan-beam geometry with straight line detectors and 3D circle cone-beam geometry with flat-plan detectors, where the simulated sinograms are generated by the open-source software “ASTRA toolbox.” We compare our method with the other existing algorithms. Our experimental results show that our new method yields the lowest root-mean-square-error (RMSE) and the highest structural-similarity (SSIM) for both reconstructed 2D and 3D fan-beam CT images.

Quantification of CH₄ coal mining emissions in Upper Silesia by passive airborne remote sensing observations with the Methane Airborne MAPper (MAMAP) instrument during the CO₂ and Methane (CoMet) campaign

Methane (CH4) is the second most important anthropogenic greenhouse gas, whose atmospheric concentration is modulated by human-induced activities, and it has a larger global warming potential than carbon dioxide (CO2). Because of its short atmospheric lifetime relative to that of CO2, the reduction of the atmospheric abundance of CH4 is an attractive target for short-term climate mitigation strategies. However, reducing the atmospheric CH4 concentration requires a reduction of its emissions and, therefore, knowledge of its sources. For this reason, the CO2 and Methane (CoMet) campaign in May and June 2018 assessed emissions of one of the largest CH4 emission hot spots in Europe, the Upper Silesian Coal Basin (USCB) in southern Poland, using top-down approaches and inventory data. In this study, we will focus on CH4 column anomalies retrieved from spectral radiance observations, which were acquired by the 1D nadir-looking passive remote sensing Methane Airborne MAPper (MAMAP) instrument, using the weighting-function-modified differential optical absorption spectroscopy (WFM-DOAS) method. The column anomalies, combined with wind lidar measurements, are inverted to cross-sectional fluxes using a mass balance approach. With the help of these fluxes, reported emissions of small clusters of coal mine ventilation shafts are then assessed. The MAMAP CH4 column observations enable an accurate assignment of observed fluxes to small clusters of ventilation shafts. CH4 fluxes are estimated for four clusters with a total of 23 ventilation shafts, which are responsible for about 40 % of the total CH4 mining emissions in the target area. The observations were made during several overflights on different days. The final average CH4 fluxes for the single clusters (or sub-clusters) range from about 1 to 9 t CH4 h−1 at the time of the campaign. The fluxes observed at one cluster during different overflights vary by as much as 50 % of the average value. Associated errors (1σ) are usually between 15 % and 59 % of the average flux, depending mainly on the prevailing wind conditions, the number of flight tracks, and the magnitude of the flux itself. Comparison to known hourly emissions, where available, shows good agreement within the uncertainties. If only emissions reported annually are available for comparison with the observations, caution is advised due to possible fluctuations in emissions during a year or even within hours. To measure emissions even more precisely and to break them down further for allocation to individual shafts in a complex source region such as the USCB, imaging remote sensing instruments are recommended.

An Improvised SIMPLS Estimator Based on MRCD-PCA Weighting Function and Its Application to Real Data

Multicollinearity often occurs when two or more predictor variables are correlated, especially for high dimensional data (HDD) where p>>n. The statistically inspired modification of the partial least squares (SIMPLS) is a very popular technique for solving a partial least squares regression problem due to its efficiency, speed, and ease of understanding. The execution of SIMPLS is based on the empirical covariance matrix of explanatory variables and response variables. Nevertheless, SIMPLS is very easily affected by outliers. In order to rectify this problem, a robust iteratively reweighted SIMPLS (RWSIMPLS) is introduced. Nonetheless, it is still not very efficient as the algorithm of RWSIMPLS is based on a weighting function that does not specify any method of identification of high leverage points (HLPs), i.e., outlying observations in the X-direction. HLPs have the most detrimental effect on the computed values of various estimates, which results in misleading conclusions about the fitted regression model. Hence, their effects need to be reduced by assigning smaller weights to them. As a solution to this problem, we propose an improvised SIMPLS based on a new weight function obtained from the MRCD-PCA diagnostic method of the identification of HLPs for HDD and name this method MRCD-PCA-RWSIMPLS. A new MRCD-PCA-RWSIMPLS diagnostic plot is also established for classifying observations into four data points, i.e., regular observations, vertical outliers, and good and bad leverage points. The numerical examples and Monte Carlo simulations signify that MRCD-PCA-RWSIMPLS offers substantial improvements over SIMPLS and RWSIMPLS. The proposed diagnostic plot is able to classify observations into correct groups. On the contrary, SIMPLS and RWSIMPLS plots fail to correctly classify observations into correct groups and show masking and swamping effects.

A component-free Lagrangian finite element formulation for large strain elastodynamics

Standard finite element formulation and implementation in solid dynamics at large strains usually relies upon and indicial-tensor Voigt notation to factorized the weighting functions and take advantage of the symmetric structure of the algebraic objects involved. In the present work, a novel component-free approach, where no reference to a basis, axes or components is made, implied or required, is adopted for the finite element formulation. Under this approach, the factorisation of the weighting function and also of the increment of the displacement field, can be performed by means of component-free operations avoiding both the use of any index notation and the subsequent reorganisation in matrix Voigt form. This new approach leads to a straightforward implementation of the formulation where only vectors and second order tensors in $${\mathbb {R}}^3$$ R 3 are required. The proposed formulation is as accurate as the standard Voigt based finite element method however is more efficient, concise, transparent and easy to implement.