K-means Clustering assisted ATR-FTIR study for rapid quality evaluation of arrowroot starch

Abstract

To solve the issues, arise in quality evaluation of arrowroot starch due to substitution and adulteration, rapid and robust k means assisted Attenuated Total Reflectance -Fourier Transform Infrared (ATR-FTIR) method is developed. Arrowroot starch, locally known as “Aerukka” or “Hulankeeriya” is the rhizomes of plants, Curcuma angustifolia Roxb. (Zingiberaceae) and Maranta arundinacea Linn. (Marantaceae). Here, ATR-FTIR technique was used to assess the arrowroot starch samples. All the samples were first analyzed in transmittance mode and the spectrum of respective samples was recorded. Resultant unique fingerprint region data was then treated with K-means clustering of machine learning. Coding for K-means clustering was executed in python. K-means is the fastest unsupervised machine learning algorithm to break down data points into groups even when very little information is available and it effectively organizing and extracting insights from complex datasets. The results showcase the effectiveness of K-means clustering as a valuable tool in analytical procedures, aiding in data-driven decision-making, and facilitating meaningful data interpretations which has distinguished all the samples in three different groups. This rapid method can be used as a routine tool to differentiate samples of arrowroot starch from its adulterants and substitutes for their quality assessment.