Open-set gas recognition: A case-study based on an electronic nose dataset

Qu, Cheng; Liu, Chuanjun; Gu, Yun; Chai, Shuiqin; Feng, Changhao; Chen, Bin

doi:10.1016/j.snb.2022.131652

Cited by 11 publications

(2 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Хоча багатовимірні сенсорні дані, одержані матрицею чутливих елементів, можуть надати більше інформації для розпізнавання газів порівняно з одноелементним сенсором, дрейф вихідних сигналів має негативний вплив на точність ідентифікації, а його врахування для багатьох чутливих елементів може значно ускладнити аналіз [18,19]. Зменшення кількості сенсорних елементів дає змогу не тільки спростити структуру штучної нюхової системи, але й покращити стабільність та ефективність детектування газу упродовж тривалого часу.…”

Section: вступunclassified

Application of Machine Learning Algorithms in Electronic Nose Technology

Olenych,

Osadchuk

2024

ELIT

View full text Add to dashboard Cite

Section: вступunclassified

Application of Machine Learning Algorithms in Electronic Nose Technology

Olenych,

Osadchuk

2024

ELIT

View full text Add to dashboard Cite

“…The flavor of the sample is exposed to the surface of the gas sensor, which results in a gas sensor response. Important information from the gas sensor response is extracted through the pattern recognition method to be classified or identified [15], [16].…”

mentioning

confidence: 99%

Design of sample display system on electronic nose for synthetic flavor classification

Barokah

Radi

Zamzami

et al. 2023

IJEECS

View full text Add to dashboard Cite

This study aimed to design a controlled sample display system on an electronic nose and test its performance for classifying synthetic flavors. There are four primary components to the electronic nose design. They are a controlled sample display system, detector, signal conditioning and preprocessing, and pattern recognition software. The sample display system consists of six vials. The sample room temperature setpoint is set to 40 ℃. The controlled sample display system has one heater and two fans to even the room temperature. The one-time data collection process consists of flushing (120 s), collecting (180 s), and purging (180 s). The samples for the performance test were synthetic flavors with four different aromas; durian, mocca, orange, and strawberry. Data analysis of gas sensor response was done through two stages; pre-treatment data processing and principal component analysis (PCA). The four samples were clearly different from others, according to the PCA results. The scores of the PC-1, PC-2, and PC-3 cumulative variance were 98.28%.

show abstract