scholarly journals Discussion of ``V-Matrix Method of Solving Statistical Inference Problems" by Vlamidir Vapnik and Rauf Izmailov

2018 ◽  
Author(s):  
Niharika Gauraha
Keyword(s):  
Conditional Probability ◽  
Optimization Problem ◽  
Probability Function ◽  
Future Research ◽  
Short Discussion ◽  
Quadratic Optimization Problem ◽  
Inference Problems ◽  
Consistent Solution ◽  
Constrained Quadratic Optimization ◽  
Conditional Probability Function

We would like to begin by stating that we have not fully understood the formulation of V-matrix conceptually. However, We are fascinated by the idea of estimation of conditional probability function without assuming any probabilistic model. In this short discussion, we would like to present that the proposed constrained quadratic optimization problem for conditional probability estimation using v-matrix based method may not have a consistent solution always. We are sure that the paper will stimulate a deeper exploration of V-matrix based methods for inference in high-dimensional problems in future research.

Discussion of the conditional-probability function for electric fields in a plasma

1984 ◽  
Vol 30 (1) ◽  
pp. 454-467 ◽  
Author(s):  
Earl W. Smith ◽  
Roland Stamm ◽  
J. Cooper
Keyword(s):  
Conditional Probability ◽  
Electric Fields ◽  
Probability Function ◽  
Conditional Probability Function

Focus on an infrequently used quantity in the context of competing risks: The conditional probability function

2018 ◽  
Vol 101 ◽  
pp. 70-81
Author(s):  
Bastien Cabarrou ◽  
Florence Dalenc ◽  
Eve Leconte ◽  
Jean-Marie Boher ◽  
Thomas Filleron
Keyword(s):  
Conditional Probability ◽  
Competing Risks ◽  
Probability Function ◽  
Conditional Probability Function

Energy-Optimal, Direct-Phase Control of Brushless Motors for Robotic Drives

2020 ◽  
Author(s):  
Amin Ghorbanpour ◽  
Hanz Richter
Keyword(s):  
Energy Consumption ◽  
Motion Control ◽  
Optimization Problem ◽  
Control Method ◽  
Bldc Motor ◽  
Brushless Motors ◽  
Energy Regeneration ◽  
Quadratic Optimization Problem ◽  
Constrained Quadratic Optimization ◽  
Ultra Capacitor

Abstract In this work, simultaneous energy regeneration and motion control for robot manipulators with brushless direct current (BLDC) motors is considered. All joints of the robot are connected to regenerative drives powered from a single ultra-capacitor. A new voltage-based control method is developed to individually command each phase of the BLDC motor. Three independent regenerative drives are interconnected in a wye configuration, and each drives a phase of the motor. The objective is to determine the control inputs for each drive to minimize energy consumption from the ultra-capacitor for a given motion task. To this end, the problem is formulated as constrained quadratic optimization problem that gives the control inputs based on the desired torque generated by a virtual controller. An experimental evaluation is performed using a pendulum actuated by a BLDC motor. It is shown that the suggested control method can accomplish the motion task and it is capable of energy regeneration. The results show a reduction of about 40% in energy consumption for the condition of the study, relative to non-regenerative case.

scholarly journals Characterization of Robust Solution Quadratically Constrained Quadratic Optimization Problem Subjected to Data Uncertainty

2021 ◽  
pp. 1-23
Author(s):  
Moussa BARRO ◽  
Satafa SANOGO ◽  
Mohamed ZONGO ◽  
Sado TRAORÉ
Keyword(s):  
Robust Optimization ◽  
Optimization Problem ◽  
Real Life ◽  
Quadratic Optimization ◽  
Data Uncertainty ◽  
Robust Solution ◽  
Quadratic Optimization Problem ◽  
Robust Duality ◽  
Constrained Quadratic Optimization ◽  
Quadratically Constrained

Robust Optimization (RO) arises in two stages of optimization, first level for maximizing over the uncertain data and second level for minimizing over the feasible set. It is the most suitable mathematical optimization procedure to solve real-life problem models. In the present work, we characterize robust solutions for both homogeneous and non-homogeneous quadratically constrained quadratic optimization problem where constraint function and cost function are uncertain. Moreover, we discuss about optimistic dual and strong robust duality of the considered uncertain quadratic optimization problem. Finally, we complete this work with an example to illustrate our solution method. Mathematics Subject Classification: (2010) 90C20 - 90C26 - 90C46-90C47 Keywords: Robust Optimization, Data Uncertainty, Quadratic Optimization Strong Duality, Robust Solution, DPJ-Convex.

scholarly journals Karakterisasi Paparan Long Term Particulate Matter di Puspiptek Serpong-Kota Tangerang Selatan

2016 ◽  
Vol 11 (1) ◽  
pp. 51 ◽  
Author(s):  
Leons Rixson ◽  
Etty Riani ◽  
Muhayatun Santoso
Keyword(s):  
Particulate Matter ◽  
Conditional Probability ◽  
Matrix Factorization ◽  
Probability Function ◽  
Positive Matrix Factorization ◽  
Positive Matrix ◽  
Conditional Probability Function

Pencemaran partikulat udara (PM10 dan PM2.5) tidak mengenal batas administrasi wilayah hal ini disebabkan bahwa pencemaran udara tidak hanya berasal dari wilayah lokal namun dapat bertransportasi dari provinsi maupun negara lain. Unsur sampel partikulat yang dikumpulkan di Serpong, Tangerang Selatan selama 24 jam periode tahun 2011 sampai dengan 2013 diidentifikasikan dengan Spekstroskopi Florensensi Sinar-X (XRF). Karekterisasi faktor dilakukan dengan menggunakan reseptor model positive matrix factorization (PMF) dan perkiraan lokasi sumber pencemar menggunakan metode Conditional Probability Function (CPF). Hasil menunjukkan rentang rata-rata konsentrasi massa PM2.5 adalah 12.63 ± 1.60 to 15.89 ± 1.70 μg/m3 sedangkan untuk PM10 berkisar 29.00 ± 3.96 to 31.04 ± 3:28 μg/m3. Multi unsur yang teridentifikasi dengan XRF adalah Al, Ca, Co, Cr, Cu, Fe, K, Mg, Mn, Na, Ni, Pb, S, Si, Ti, V dan Zn. Karakterisasi partikulat halus (PM2.5) teridentifikasi 5 faktor yaitu industri peleburan logam Pb (9.61%), debu tanah (17%), campuran industri peleburan logam dan garam laut (13.02%), trasnportasi (44.36%) serta pembakaran biomassa (22.58%).Kata kunci : polusi udara, spektroskopi XRF, sumber pembagian, fungsi probabilitasbersyarat

Two-Stage Feature Selection with Unsupervised Second Stage

2018 ◽  
Vol 27 (07) ◽  
pp. 1860014
Author(s):  
Ke Xu ◽  
Crystal Maung ◽  
Hiromasa Arai ◽  
Haim Schweitzer
Keyword(s):  
Feature Selection ◽  
Optimization Problem ◽  
Complexity Analysis ◽  
Feature Selection Problem ◽  
Quadratic Optimization Problem ◽  
Second Stage ◽  
Dimensionality Reduction Technique ◽  
Constrained Quadratic Optimization ◽  
Two Stages ◽  
Simple Filter

Feature selection is a common dimensionality reduction technique of fundamental importance in big data. A common approach for reducing the running time of feature selection is to perform it in two stages. In the first stage a fast and simple filter is applied to select good candidates. The number of candidates is further reduced in the second stage by an accurate algorithm that may run significantly slower. There are two main variants of feature selection: unsupervised and supervised. In the supervised variant features are selected for predicting labels, while the unsupervised variant does not use labels at all. We describe a general framework that can use an arbitrary off-the-shelf unsupervised algorithm for the second stage. The algorithm is applied to the selection obtained in the first stage weighted appropriately. Our main technical result is a method for calculating weights for the columns that need to be selected in the second stage. We show that these weights can be computed as the solution to a constrained quadratic optimization problem. The solution is deterministic, and improves on previously published studies that use probabilistic ideas to compute similar weights. To the best of our knowledge our approach is the first technique for converting a supervised feature selection problem into an unsupervised problem. Complexity analysis shows that the proposed technique is very fast, can be implemented in a single pass over the data, and can take advantage of data sparsity. Experimental results show that the accuracy of the proposed method is comparable to that of much slower techniques.

Sign in / Sign up

Export Citation Format

Share Document