Date of Publication

2009

Document Type

Master's Thesis

Degree Name

Master of Science in Computer Science

Subject Categories

Computer Sciences

College

College of Computer Studies

Department/Unit

Computer Science

Thesis Adviser

Rigan P. Ap-Apid

Abstract/Summary

Biochemical networks contain valuable information that can be extracted for analysis. One important component is to determine the dynamics of a biochemical network and its structure. The representation of the biochemical network is modeled using the S-system model. This model needs its parameter to be estimated. A parameter estimation algorithm is used to estimate these parameters. There are different papers for parameter estimation algorithms available. Each one informs the result and capability of the algorithm. However, these are tested on different biochemical networks and the environmental configurations are most likely different. These may lead to difficulty in comparing to other algorithms. This study is part of a project that benchmarks other parameter estimation algorithm on uniform test data. The aim of this study is to focus on implementing and then benchmarking the S-tree based algorithm with the same uniform test data. The algorithm is tested on different biochemical networks. And, the performance of the algorithm is then analyzed. The performed analysis may provide a guide in exploring improvements for the algorithm and proceed to the implementation and benchmarking of the same test data as well. A test on inclusion of balancing of error and complexity is done to determine the improvement of the algorithm. In addition, the decoupling of the data is taken advantage and tested for effectiveness. Keywords: Benchmarking, Performance Analysis, Parameter Estimation, Biochemical Network, S-System, Simulation, S-tree, Genetic Programming, Balancing of Error and Complexity, Decoupling Biochemical networks contain valuable information that can be extracted for analysis. One important component is to determine the dynamics of a biochemical network and its structure. The representation of the biochemical network is modeled using the S-system model. This model needs its parameter to be estimated. A parameter estimation algorithm is used to estimate these parameters. There are different papers for parameter estimation algorithms available. Each one informs the result and capability of the algorithm. However, these are tested on different biochemical networks and the environmental configurations are most likely different. These may lead to difficulty in comparing to other algorithms. This study is part of a project that benchmarks other parameter estimation algorithm on uniform test data. The aim of this study is to focus on implementing and then benchmarking the S-tree based algorithm with the same uniform test data. The algorithm is tested on different biochemical networks. And, the performance of the algorithm is then analyzed. The performed analysis may provide a guide in exploring improvements for the algorithm and proceed to the implementation and benchmarking of the same test data as well. A test on inclusion of balancing of error and complexity is done to determine the improvement of the algorithm. In addition, the decoupling of the data is taken advantage and tested for effectiveness. Keywords: Benchmarking, Performance Analysis, Parameter Estimation, Biochemical Network, S-System, Simulation, S-tree, Genetic Programming, Balancing of Error and Complexity, Decoupling.

Abstract Format

html

Language

English

Format

Print

Accession Number

CDTG004522

Shelf Location

Archives, The Learning Commons, 12F Henry Sy Sr. Hall

Physical Description

ii, 49 leaves ; 28 cm.

Keywords

Benchmarking; Algorithms; Algorism; Simulation; Genetic programming (Computer science)

Upload Full Text

wf_yes

Download

Share

COinS