Evaluation and categorization of raw cassava log quality for structural applications

ABSTRACT

Cassava used in various processing industries needs new tools and technology to ensure it's the final quality Cassava product based on the component exists in the raw root to make the final product competitive. Modifying industry currently uses the traditional wet chemical method for ingredient identification as per Indian Standard assessment system developed in 1978. It was expensive, take prolonged testing time, need skilful worker involvement and chance of toxic byproduct formation. This paper focuses on the pre-determined raw material quality feeding for suitable starch industry through raw Cassava categorization based on its active constituent presence to meet the final quality of modified Cassava starch by integrating Fourier Transform Infra-Red spectroscopy (FTIR) with Partial Least Square (PLS) algorithm. The classification of raw cassava logs based on its ingredient concentration (ash and moisture) implemented through Support Vector Machine algorithm according to the industrial standard requirements needed by the third party users for guaranteed final quality product from raw materials observation.

Keywords:
Fourier Transform Infrared Red (FT-IR); Support Vector Machine (SVM); Partial Least Squares (PLS); Root Mean Square Error of Prediction (RMSEP); Relative Percent Difference (RPD)