dos Reis M. M., Ferreira M. M. C., Sarmento Silene B. S., “A methodological multi-way analysis of Cassava Starch properties”.  Copenhagen, Denmark, 19-23/08/2001: 7^th Scandinavian Symposium on Chemometrics (SSC7), Book of Abstracts, A76 (2001). Poster P36.

A methodological multi-way analysis of Cassava
Starch properties

Marlon M. Reis, marlon@iqm.unicamp.br, Chemistry Institute - UNICAMP. Cidade Universitária Zeferino
             Vaz, s/n, Campinas, Brazil.
Márcia M. C. Ferreira, marcia@iqm.unicamp.br, Universidade Estadual de Campinas, Instituto de Química.
            Campinas, SP, 13083-970, Brazil
Silene B. S. Sarmento, Department of Agroindustry, Food and Nutrition, "Escola Superior de Agricultura - Luiz
          de Queiroz" Universidade de São Paulo, Piracicaba, Brazil

Keywords: multiway exploratory analysis, Cassava starch, constrained, TUCKER

The original methods proposed by Ledyard R. Tucker during the 1960s for Multi-Way
Analysis present the rotational freedom problem, making the interpretation of its results rather
difficult to be carried out. With the goal of making the multi-way data analysis more
straightforward, this work uses a methodology of extracting meaningful information from the
data set. The present methodology is based on the decomposition of data set in 3-way blocks
by using Constrained Tucker Model. The aim of using this approach is to keep in one block
all the similar information about data properties. The decomposition used is due to a
Constrained Tucket Model, where the core array has some of its elements fixed to zero.

This work deals with a data set formed by the properties of four cassava cultivars, harvested at
different ages during the usual harvesting of cultivars for industrial usage (age-properties-
cultivars). The formulation of the Constrained Tucker Model considered independence among
blocks. The inertia function, which gave information about how the data slices are described
by the core slice, was fundamental for the final adjustment of the Constrained Tucker Model.
This methodology is interesting since the vectors on the A and B modes, which show the
correlation between properties and age, are directly related in one block, making its analysis
quite easy. Although the correlation among the considered properties, the starch structure, the
age stage and the cultivars varieties presents itself as a complex puzzle, the three-way analysis
carried out showed to be able to provide useful information about the data, helping to choose
the best harvesting period, considering the starch's properties which are important for
industry

The authors acknowledge the financial support from FAPESP  for carrying out this work.
 
 
 
 

                                                                     A76