Alternative sensory eval - DA
Quoc Cuong Nguyen
2022-11-14
Clean
rm(list = ls()); graphics.off()Set R packages path and load packages
.libPaths("C:/Rpack")
library(SensoMineR)## Loading required package: FactoMineRlibrary(openxlsx)
library(FactoMineR)Import data
bread_qda <- read.xlsx("DAdata.xlsx", sheet = 1)
head(bread_qda)## assessor product replicate Hue Intensity Whiteness Saltiness Spice Bitterness
## 1 A8 P1 R1 3 3 8 7 1 1
## 2 A8 P1 R2 3 2 5 5 1 1
## 3 A8 P2 R1 3 3 8 7 1 1
## 4 A8 P2 R2 2 2 8 6 1 1
## 5 A8 P3 R1 1 1 9 7 3 1
## 6 A8 P3 R2 2 1 9 6 3 1
## Hardness Crispness Fatness
## 1 3 6 6
## 2 5 4 7
## 3 2 5 5
## 4 7 7 6
## 5 3 7 5
## 6 2 5 5Run analyses
Panel performance
ANOVA model
qda <- as.data.frame(bread_qda)
res.pp <- panelperf(qda, firstvar=4, formul="~product+assessor+replicate+ product:assessor+product:replicate+assessor:replicate")
p.value <- res.pp$p.value
p.value## product assessor replicate
## Hue 1.594292e-13 1.641190e-04 0.8667318
## Intensity 6.859327e-16 1.428904e-05 0.7910052
## Whiteness 8.911465e-15 3.319416e-06 0.8619209
## Saltiness 9.358085e-02 2.739611e-05 0.2988723
## Spice 1.877678e-06 7.952131e-13 0.3200119
## Bitterness 1.772001e-01 7.739878e-10 0.1945945
## Hardness 5.951338e-08 6.928288e-05 0.1500957
## Crispness 1.434990e-04 1.956800e-05 0.6682310
## Fatness 1.134765e-02 3.057369e-06 0.8182805
## product:assessor product:replicate assessor:replicate
## Hue 6.382259e-07 0.4596379 0.55454349
## Intensity 2.047463e-03 0.3998635 0.03276646
## Whiteness 1.590794e-01 0.7276830 0.03307057
## Saltiness 1.607340e-02 0.7490344 0.26589052
## Spice 1.429352e-04 0.5305475 0.88122935
## Bitterness 6.098997e-05 0.1486547 0.01579670
## Hardness 1.080251e-01 0.5466320 0.49226407
## Crispness 1.684461e-02 0.3822134 0.99130389
## Fatness 6.327374e-03 0.5534993 0.74357779Plot and save ANOVA table
windows(10,10)
coltable(p.value[order(p.value[,1]),], col.lower="gray")
# savePlot("ANOVA tab", type = "tif")PCA
Generate averaged data
res.decat <- decat(qda, formul="~product+assessor", firstvar=4,
lastvar=ncol(qda), graph=F)
qda_m <- res.decat$adjmean
qda_m## Hue Intensity Whiteness Saltiness Spice Bitterness Hardness Crispness
## P1 4.4375 4.4375 5.0000 4.8125 2.1250 1.4375 5.0000 5.8125
## P2 2.3750 2.6875 7.2500 5.3750 1.8750 1.3125 4.7500 5.1875
## P3 2.0625 2.0000 7.8750 5.1250 3.6875 1.3125 3.0000 4.3750
## P4 6.5000 6.6250 3.1250 5.1250 2.3750 1.9375 7.8125 7.3750
## P5 2.6875 2.9375 6.6250 6.2500 3.6250 1.8750 5.8125 6.7500
## P6 7.8750 7.2500 2.5625 4.1250 5.5625 2.0625 4.8750 5.6250
## Fatness
## P1 5.5625
## P2 4.7500
## P3 4.0625
## P4 6.2500
## P5 4.5000
## P6 5.2500Apply PCA
res.pca <- PCA(qda_m, graph = FALSE)
res.dimdesc <- dimdesc(res.pca)Plot PCA results
windows(10,10)
plot(res.pca, axes = c(1,2), choix = "ind", col.var = "blue", graph.type="classic")
# savePlot("PCA scores", type = "tif")
windows(10,10)
plot(res.pca, axes = c(1,2), choix = "var", col.var = "red", graph.type="classic")
# savePlot("PCA loadings", type = "tif")Spider plot
Install package and set up
library(fmsb)
RNGkind("Mersenne-Twister")
set.seed(123)
maxmin <- t(data.frame(rep(9,ncol(qda_m)), rep(1,ncol(qda_m))))
rownames(maxmin) <- c("max", "min")
colnames(maxmin) <- colnames(qda_m)
dat <- as.data.frame(rbind(maxmin,qda_m))
cols <- c("red", "lightblue", "forestgreen", "purple",
"hotpink", "chocolate4", "orange", "steelblue",
"grey", "green", "khaki", "black")Plot spider-plot
windows(10,10)
radarchart(dat, axistype=0, seg=4, plty=1, vlabels= colnames(dat),
title="Spider plot", vlcex=1,
pcol = cols, plwd = 2)
legend("topright",
legend = rownames(dat)[3:nrow(dat)],
bty = "n", pch = 20, col = cols,
text.col = cols, pt.cex = 2)
# savePlot("Spider plot", type = "tif")