En nuestra última edición

Low-cost meteorological platform based on zigbee technology

Computer- assisted analysis for e-portfolio learning strategy

Smart home control by voice using neural networks

Application software as global marketing strategy

Methodology used in the design and construction of a computer numerical control machine

Exergy analysis for a furnace in the production of panela

Reflection on the requirements specification in a university context

Digital social networks: an approach to risk management information systems

Analysis of the current system of a clay block dryer and an improvement proposal

Algorithms and systems based on patterns of stem cell and artificial cloning for automation and control in an assembly operation in metalworking processes

Development of a software to apply zoometria in bovine cattle

Design and simulation of a prototype of machine for signaling of ways, automated through a mechanical system

Modelation of modal election for urban public transport based on the introduction of a new mode

Evaluation of resistance to flexion in glazed ceramic tile in terms of ice and thaw cycles

A look at the onto- epistemic fundamentals of the career of system engineering as a basis for rethinking the social impact in times of changing

Structural analysis of a pressing machine for production of solid brick for small craft industries of ceramic materials in Ocaña, Norte de Santander, and in the region

Identification of vegetable coverages from satellite images over the middle and lower basin of the Catatumbo river

Aplication of lattice boltzmann equation method to the bidimensional simulation of typical fluids mechanic problem

The historical references of professional. development case study of teachers teaching in the basic primary education in Colombia

Critical thinking of students using facebook like new technology learning

Resumen

 

FAST VARIABLE SELECTION FOR MASS SPECTROMETRY ELECTRONIC NOSE APPLICATIONS

RAPIDA SELECCION DE VARIABLES PARA APLICACIONES DE NARICES ELECTRONICAS BASADAS EN ESPECTROMETRIA DE MASAS
PhD. Oscar Eduardo Gualdrón Guerrero.
PhD. Ivaldo Torres Chávez.

Abstract: High dimensionality is inherent to MS-based electronic nose applications where hundreds of variables per measurement (m/z fragments) ¾ a significant number of them being highly correlated or noisy ¾ are available. Feature selection is, therefore, an
unavoidable pre-processing step if robust and parsimonious pattern classification models are to be developed. In this article, a new strategy for feature selection has been introduced and its good performance demonstrated using two MS e-nose databases. The feature selection is conducted in three steps. The first two steps are aimed at removing noisy, non-informative and highly collinear features (i.e., redundant), respectively. These two steps are computationally inexpensive and allow for dramatically reducing the number of variables (near 80% of initially available features are eliminated after the second step). The third step makes use of a stochastic variable selection method (simulated annealing) to further reduce the number of variables. For example, applying the method to an Iberian ham database has resulted in the number of features being reduced from 209 down to 14. Using the surviving m/z fragments, a fuzzy ARTMAP classifier was able to sort ham samples according to producer and quality (11-category classification) with a 97.24% success rate. The whole feature selection process runs in a few minutes in a Pentium IV PC platform.

Resumen: Una alta dimensionalidad es inherente en aplicaciones de narices electrónicas basadas en MS, donde se pueden encontrar cientos de variables por medida, un número significativo de ellas proporcionan ruido o una alta correlación entre ellas. En este articulo, una nueva estrategia de selección de variables es desarrollada con buenos resultados usando dos bases de datos de narices electrónicas basadas en MS. El proceso se realizo en tres pasos. En los dos primeros pasos el objetivo es eliminar ruido e información altamente colineal (redundancia), respectivamente. El tercer paso se utiliza el método de selección estocástico (simulated annealing) para reducir significativamente el número de variables. El proceso de selección total se ejecuto en pocos minutos en una plataforma Pentium IV.

Keywords : Feature selection, Mass spectrometry, Electronic nose, Simulated annealing, Neural networks, Iberian ham.