Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Jun 27, 2026

Tea Aroma Analysis Based on Solvent-Assisted Flavor Evaporation Enrichment
04:36

Tea Aroma Analysis Based on Solvent-Assisted Flavor Evaporation Enrichment

Published on: May 26, 2023

Linking Tea Aroma Chemistry to Quality Grades via a Single MOS Gas Sensor: Classical Machine Learning vs. Deep

Ahmet Turan Tasdemir1, Erkan Caner Ozkat2,3, Gozde Yalcin Ozkat4

  • 1Department of Mechanical Engineering, Institute of Graduate Studies, Recep Tayyip Erdogan University, Rize 53100, Türkiye.

Sensors (Basel, Switzerland)
|June 26, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Design, synthesis, and computational studies of benzimidazole derivatives as new antitubercular agents.

Journal of biomolecular structure & dynamics·2022
See all related articles

A new method uses a single low-cost gas sensor and waveform analysis to accurately determine black tea quality based on aroma chemistry. This approach surpasses traditional methods by analyzing the temporal release of volatile compounds for precise grading.

Area of Science:

  • Food Science
  • Analytical Chemistry
  • Sensor Technology

Background:

  • Black tea quality is determined by aroma compounds like terpene alcohols and aldehydes.
  • Traditional electronic noses compress volatile organic compound (VOC) signals, losing crucial release kinetic information.
  • Assessing tea grade requires analyzing both the abundance and release patterns of aroma chemistry.

Purpose of the Study:

  • To investigate if a single metal-oxide-semiconductor (MOS) gas sensor can identify grade-defining aroma chemistry in black tea.
  • To evaluate if waveform-level deep learning models can effectively utilize sensor data for tea quality assessment.
  • To compare the performance of traditional feature-based machine learning with deep learning on raw sensor waveform data.

Main Methods:

Keywords:
MOS gas sensorblack tea qualitydeep learningelectronic nosevolatile organic compounds

More Related Videos

Integration of Animal Behavioral Assessment and Convolutional Neural Network to Study Wasabi-Alcohol Taste-Smell Interaction
06:19

Integration of Animal Behavioral Assessment and Convolutional Neural Network to Study Wasabi-Alcohol Taste-Smell Interaction

Published on: August 16, 2024

Related Experiment Videos

Last Updated: Jun 27, 2026

Tea Aroma Analysis Based on Solvent-Assisted Flavor Evaporation Enrichment
04:36

Tea Aroma Analysis Based on Solvent-Assisted Flavor Evaporation Enrichment

Published on: May 26, 2023

Integration of Animal Behavioral Assessment and Convolutional Neural Network to Study Wasabi-Alcohol Taste-Smell Interaction
06:19

Integration of Animal Behavioral Assessment and Convolutional Neural Network to Study Wasabi-Alcohol Taste-Smell Interaction

Published on: August 16, 2024

  • A portable electronic nose with a Bosch BME688 sensor recorded time-series data (temperature, humidity, pressure, gas resistance) from 16 Turkish black teas across three grades.
  • Two data representations were compared: statistical feature extraction (principal components) for classical classifiers and raw sensor waveforms for a deep learning model (MS-CNN-Attention).
  • Model performance was evaluated using product-grouped cross-validation and an F1-macro score, focusing on classification accuracy at the product level.

    • Main Results:

    • The deep learning model (MS-CNN-Attention) achieved a significantly higher F1-macro score (0.811) compared to the feature-based MLP model (0.624), a 30% improvement.
    • The deep learning approach correctly graded 14 out of 16 products, demonstrating superior performance, especially for the medium quality grade (F1: 0.52 to 0.79).
    • The study highlighted that waveform-level analysis, unlike summary statistics, preserves the critical release-kinetic signal essential for distinguishing tea grades.

    Conclusions:

    • A single programmable MOS sensor, when used as a thermal-desorption profiler with waveform-level modeling, can effectively screen black tea quality.
    • Deep learning models analyzing raw sensor waveforms offer a substantial advantage over traditional feature-based methods for aroma chemistry analysis.
    • This pairing provides a rapid, non-destructive, and cost-effective method for aroma-chemistry-based tea quality assessment and batch screening.