A local weighted nearest neighbor algorithm and a weighted and constrained least-squared method for mixed odor...

Area of Science:

• Chemometrics
• Sensor Technology
• Analytical Chemistry

Background:

• Electronic nose systems typically identify whole odors by comparison to databases.
• Direct analysis of individual odor components is often more practical but challenging.
• Existing methods lack precision in determining specific odorant contributions within mixtures.

Purpose of the Study:

• To develop novel methods for direct odor component analysis in electronic nose systems.
• To accurately identify individual odorants within a mixture.
• To quantify the concentration of each identified odor component.

Main Methods:

• A K-nearest neighbor (KNN)-based local weighted nearest neighbor (LWNN) algorithm for odor component classification.
• Odor training data categorized into groups represented by centroids.
• A weighted and constrained least-squares (WCLS) method for estimating component concentrations via regression models.

Main Results:

• The LWNN algorithm successfully classified mixed odors with varying mixing ratios.
• The WCLS method provided accurate estimations of individual component concentrations.
• The proposed methods demonstrate effectiveness in direct odor component analysis.

Conclusions:

• The developed LWNN and WCLS methods offer a significant advancement for electronic nose systems.
• These techniques enable direct identification and quantification of odor components, moving beyond simple odor identification.
• The findings pave the way for more sophisticated and practical applications of electronic noses in various fields.