직/편성물 염색

Neuro-Fuzzy Modelling of Spectroscopic Data. Part B - Dye Concentration Prediction

  • 출판일1997.09
  • 저자
  • 서지사항
  • 등록일 2016.11.02
  • 조회수 294
Researchers investigated the neuro-fuzzy modeling of spectroscopic data using wavelength and absorbance as model inputs and concentration as the output in the prediction application. Application of an adaptive neuro-fuzzy inference system used a training data set. The data set consisted of 342 absorbance values in a wavelength range of 400-580 nanometers for 18 calibration samples dyed with red or yellow leather dyes, their dye concentration values, and 19 wavelength values. Testing of predictive ability of the model used two tests sets, each consisting of 23 samples with dye concentrations spread over the range of 0-900 milligrams per liter. Results indicated that the model required five membership functions for two imputs, wavelength and absorbance, to achieve a good fit in the model. The prediction error for the model increased at higher concentrations due to dye molecule interactions. 11 refs.
이전글 다음글 목록