Skip to main content

Development of data mining technique for spectroscopic analysis results of biochemical samples

Primary supervisor

Chunyang Chen

In situ analysis of complex biochemical metrices, such as microbial fermentation products, has drawn substantial research interests in recent years. Compared to the commonly used chemical analysis technique, spectroscopic analysis has great potential for this purpose due to its advantages of being rapid, contactless, lossless, and solvent-free. The major obstacle for the further application of the spectroscopic technique in the analysis of biochemical samples is the lack of customized data mining methods for the highly complicated spectral signal of the organic mixtures. The current project is aimed at the optimization and development of signal processing, statistical, machine learning, and deep learning methods based on the features of the time series-like 2D spectral data of samples generated from the industrial fermentation process to aid the product quality and safety assessment. The resulted artificial intelligence-assisted spectroscopic technique would significantly promote the reliability level of in situ analysis of complex biochemical metrices.

Student cohort

Double Semester

Required knowledge

  • Machine Learning & Deep Learning
  • Python programming
  • Knowledge about time-series data analysis will be a big plus