The exploration of the potency of ftir spectroscopy and chemometrics for extract bioactivity prediction and rapid analysis of bioactive content of medicinal plants

Abstract

An infrared spectra is a characteristic representation for a chemistry system. There are many researches have been conducted for the interpretation supported by chemometric method which aimed for metabolite fingerprinting. Nowadays, the application of IR spectroscopy and chemometric for quantitative analysis of mixture system is revealed to be increasing because of ease of use, speed, reftabtfty, s«rnpJe size, and catitrafon stability. The paper is about the use of FTIR spectroscopy technique and chemometric for bidding a model of bioactrvrty prediction of jatibelanda extract (extract inhibition to the lipase enzyme activity} and generating calibration model tor determination gingero! in powder and ginger extract Both of them were made by three steps, i.e. the data reduction for reducing the complexity of data structure of infrared spectra, the build of' cafibrabofi model, and the model validation. The data reduction technique consisted of principal component analysis (PCA) dan fragmented regression. The generating calibration model and validation method used artificial neural network (ANN) dan Bayes approach. The ANN bioactvity model can make wefi-mannered of grouping of inhibition characteristic of jatibelanda extract and the calibration model of gingeroi by the Bayes approach resulted high precise prediction