TITLE:
Volatile Compounds Selection via Quantile Correlation and Composite Quantile Correlation: A Whiting Case Study
AUTHORS:
Ibrahim Sidi Zakari, Assi N’guessan, Alexandre Dehaut, Guillaume Duflos
KEYWORDS:
Volatile Compounds, Freshness and Spoilage Indices, Quantile Correlation, Composite Quantile Correlation, Sure Independence Screening
JOURNAL NAME:
Open Journal of Statistics,
Vol.6 No.6,
November
14,
2016
ABSTRACT: The freshness and quality indices of
whiting (Merlangius merlangus) influenced by a large number of chemical
volatile compounds, are here analyzed in order to select the most relevant
compounds as predictors for these indices. The selection process was performed
by means of recent statistical variable selection methods, namely robust
model-free feature screening, based on quantile correlation and composite
quantile correlation. On the one hand, compounds 2-Methyl-1-butanol,
3-Methyl-1-butanol, Ethanol, Trimethylamine, 3-Methyl butanal,
2-Methyl-1-propanol, Ethylacetate, 1-Butanol and 2,3-Butanedione were
identified as major predictors for the freshness index and on the other hand,
compounds 3-Methyl-1-butanol, 2-Methyl-1- butanol, Ethanol, 3-Methyl butanal,
3-Hydroxy-2-butanone, 1-Butanol, 2,3-Butane- dione, 3-Pentanol, 3-Pentanone and
2-Methyl-1-propanol were identified as major predictors for the quality index.