On semiseparable kernels and efficient implementation for regularized system identification and function estimation

Automatica ◽  
2021 ◽  
Vol 132 ◽  
pp. 109682
Author(s):  
Tianshi Chen ◽  
Martin S. Andersen
Keyword(s):  
System Identification ◽  
Efficient Implementation ◽  
Function Estimation ◽  
And Function

scholarly journals Kernel methods in system identification, machine learning and function estimation: A survey

Automatica ◽  
2014 ◽  
Vol 50 (3) ◽  
pp. 657-682 ◽  
Author(s):  
Gianluigi Pillonetto ◽  
Francesco Dinuzzo ◽  
Tianshi Chen ◽  
Giuseppe De Nicolao ◽  
Lennart Ljung
Keyword(s):  
Machine Learning ◽  
System Identification ◽  
Kernel Methods ◽  
Function Estimation ◽  
And Function

scholarly journals Near-Linear Time Local Polynomial Nonparametric Estimation with Box Kernels

2021 ◽  
Author(s):  
Yining Wang ◽  
Yi Wu ◽  
Simon S. Du
Keyword(s):  
Operations Management ◽  
Polynomial Regression ◽  
Linear Time ◽  
Big Data Analytics ◽  
Function Estimation ◽  
Local Polynomial ◽  
Efficiency And Effectiveness ◽  
Exploratory Data ◽  
And Function ◽  
Novel Applications

Summary of Contribution: Big data analytics has become essential for modern operations research and operations management applications. Statistics methods, such as nonparametric density and function estimation, play important roles in predictive and exploratory data analysis for economics and operations management problems. In this paper, we concentrate on efficiently computing local polynomial regression estimates. We significantly accelerate the computation of such local polynomial estimates by novel applications of multidimensional binary indexed trees ( Fenwick 1994 ) and lazy memory allocation via hashing. Both time and space complexity of our proposed algorithm are nearly linear in the number of inputs. Simulation results confirm the efficiency and effectiveness of our proposed methods.

scholarly journals Mixed Methods in Tactical Analysis Through Polar Coordinates and Function Estimation: The Transition Play in ACB Basketball

2021 ◽  
Vol 3 ◽  
Author(s):  
José Luis Pastrana-Brincones ◽  
Belén Troyano-Gallegos ◽  
Juan Pablo Morillo-Baro ◽  
Raimundo López de Vinuesa-Piote ◽  
Juan Antonio Vázquez-Diz ◽  
...  
Keyword(s):  
Ad Hoc ◽  
High Reliability ◽  
Reliability And Validity ◽  
Correlation Coefficients ◽  
Polar Coordinates ◽  
Function Estimation ◽  
Mating Behaviors ◽  
The One ◽  
High Level ◽  
And Function

Nowadays, getting advantageous offensive situations in high-level basketball is being increasingly harder, so taking advantage of any situation in the game since the team has the ball is essential to be competitive. Therefore, the goal to achieve in this study is to evaluate using a mixed method strategy the behaviors happening in the application of the technical–tactical means performed in the transition play of professional basketball in Spain. An ad hoc observation tool made of 11 criteria and 83 exhaustive and mutually exclusive categories (E/ME) has been designed and validated by means of data quality and generalizability analyses. Indexes obtained show high reliability and validity allowing the proposed actions to be recorded (correlation coefficients are above 0.95 and generalizability coefficients are above 0.90 in all cases). A total number of 128 situations corresponding to eight games of Unicaja de Málaga in the Endesa League in the 18/19 season were observed with the Hoisan software. The analysis of the relationships among behaviors was performed using the polar coordinates technique where the one-on-one initiation, outside the zone, has been used as focal behavior. The estimation of the functions representing the vectors has also been performed to model the best fit that estimates, starting from a focal category, the relationship among this focal behavior and the rest of the mating behaviors for possible future observations. The results show significant relationships between the selected focal behavior and the mating behaviors, showing indications of behaviors allowing tactical interpretation of the game and the definition of intervention programs to improve the performance of the team.

A Novel Methodology for Combined Parameter and Function Estimation Problems

2009 ◽  
Author(s):  
Ali Hakkaki-Fard ◽  
Hosein Molavi ◽  
Ramin K. Rahmani
Keyword(s):  
Measurement Errors ◽  
Present Method ◽  
Adjoint Problem ◽  
Limiting Factor ◽  
Function Estimation ◽  
Estimation Problems ◽  
Convergence Performance ◽  
Complete Set ◽  
And Function ◽  
Correlated Parameters

This article presents a novel methodology, which is highly efficient and simple to implement, for simultaneous retrieval of a complete set of thermal coefficients in combined parameter and function estimation problems. Moreover, the effect of correlated parameters on convergence performance is examined. The present methodology is a combination of two different methods: The Conjugate Gradient Method with Adjoint Problem (CGMAP) and Box-Kanemasu method (BKM). The methodology uses the benefit of CGMAP in handling function estimation problems and BKM for parameter estimation problems. One of the unique features about the present method is that the correlation among the separate unknowns does not behave as a limiting factor to the convergence of the problem. Numerical experiments using measurement errors are performed to verify the proposed method in solving the combined parameter and function estimation problems. The obtained results show that the combined procedure can efficiently and reliably estimate the values of the thermal coefficients.

Sign in / Sign up

Export Citation Format

Share Document