Abstract

Journal of Near Infrared Spectroscopy
Volume 17 Issue 2, Pages 69–76 (2009)
doi: 10.1255/jnirs.830

A chemometrics approach for distinguishing between beers using near infrared spectroscopy

Hua Li,^a Yutaka Takahashi,^a Masanori Kumagai,^b Kazuhiko Fujiwara,^a Ryoei Kikuchi,^a Noboru Yoshimura,^a Toshio Amano,^c Jinming Lin^d and Nobuaki Ogawa^a,*
aFaculty of Engineering and Resource Science, Akita University, Tegata Gakuencho, Akita 010-8502, Japan. E-mail: ogawa@ipc.akita-u.ac.jp
^bInstitute for Food and Brewing, Akita Prefectural Agriculture, Forestry and Fisheries Research Center, Arayamachi, Akita 010-1623, Japan
^cOPT Research. Inc., 5-1-3, Minami-Aoyama, Mimato-ku, Tokyo 107-0062, Japan
^dDepartment of Chemistry, Tsinghua University, Beijing 100084, China

Thirty-eight beers from different producing areas and/or makers were distinguished by principal component analysis (PCA) of the near infrared (NIR) spectra acquired by a portable NIR spectrometer. Classsification of Akita beers: beers locally produced in Akita prefecture, Japan, from other famous brand beers could be successfully performed, especially when the PCA was calculated on the standard normal variate (SNV) spectra. The classification equations use information related to water and CH₂ absorption that reflected the differences in chemical composition of beers due to different production processes. In addition, the compositions of total polyphenol and total nitrogen were estimated from NIR spectra by multiple linear regression (MLR). This study showed that NIR spectroscopy is promising for beer quality evaluation, both for identifying multifarious beers including Akita beers using PCA and for rapid in-line quality control and inspection for beer production using the quantitative MLR analysis.

Keywords: near infrared spectroscopy, portable spectrometer, principal component analysis, multiple linear regression, distinction, beer

---

Full-text article (646 kB) (subscribers only)

Buy article on-line for £20 (get immediate access)

Alerting Service

 RSS Feed

Permalink: http://dx.doi.org/10.1255/jnirs.830
QR Code (what is this?):

---