Advances in predicting human olfactory perception: from data acquisition to computational models.

Area of Science:

• Odor analysis and computational chemistry
• Sensor technology and materials science
• Machine learning and artificial intelligence

Background:

• Accurate odor prediction requires sophisticated data processing and advanced sensor capabilities.
• Gas sensing technologies are crucial for identifying gas mixtures and providing response signals.
• Understanding human olfactory perception is a key goal in odor research.

Purpose of the Study:

• To explore recent advances in gas sensing technologies for human olfactory perception.
• To summarize olfactory perception databases and fundamental sensing principles.
• To connect odor sensing technologies with suitable machine learning algorithms.

Main Methods:

• Review of olfactory perception databases.
• Introduction to sensing principles of gas chromatography-mass spectrometry, metal-oxide semiconductor, optical, and electrochemical sensors.
• Integration of odor sensing technologies with machine learning algorithms (ANN, RF, KNN, SVM, ELM, GBDT, DT).

Main Results:

• Identification of key gas sensing technologies relevant to molecular odor prediction.
• Overview of fundamental principles for various sensor types.
• Mapping of sensing technologies to appropriate machine learning algorithms for odor analysis.

Conclusions:

• Machine learning is essential for processing complex odor data.
• Advanced gas sensors are critical for accurate odor detection and characterization.
• Future integration of ML and sensing will significantly advance olfactory research and understanding of human sensory mechanisms.