Add files via upload

Author: guokai8

vignettes/o2plsda.Rmd

@@ -40,12 +40,6 @@ The relation between $T$ and $U$ makes the joint part the joint part: $U = TB_U
 In order to avoid overfitting of the model, the optimal number of latent variables for each model structure was estimated using group-balanced Monte Carlo cross-validation (MCCV). The package could use the group information when we select the best parameters with cross-validation. In cross-validation (CV) one minimizes a certain measure of error over some parameters that should be determined a priori. Here, we have three parameters: $(nc, nx, ny)$. A popular measure is the prediction error $||Y - \hat{Y}||$, where $\hat{Y}$ is a prediction of $Y$. In our case the O2PLS method is symmetric in $X$ and $Y$, so we minimize the sum of the prediction errors: 
 $||X - \hat{X}||+||Y - \hat{Y}||$. 
 
-And we also calculate the the average $Q^2$ values:
-
-$Q^2$ = 1 - $err$ / $Var_{total}$;    
-
-$err$ = $Var_{expected}$ - $Var_{estimated}$;   
-
 Here $nc$ should be a positive integer, and $nx$ and $ny$ should be non-negative. The 'best' integers are then the minimizers of the prediction error.
 
 The O2PLS-DA analysis was performed as described by Bylesjö et al. (2007); briefly, the O2PLS predictive variation [$TW^\top$, $UC^\top$] was used for a subsequent O2PLS-DA analysis. The Variable Importance in the Projection (VIP) value was calculated as a weighted sum of the squared correlations between the OPLS-DA components and the original variable.
@@ -81,39 +75,19 @@ set.seed(123)
 ## ncores : parallel paramaters for large datasets
 cv <- o2cv(X,Y,1:5,1:3,1:3, group = group, nr_folds = 10)
 #####################################
-# The best parameters are nc =  5 , nx =  3 , ny =  3 
+#The best parameters are nc = 1, nx = 2, ny = 3
 #####################################
-# The Qxy is  0.08222935  and the RMSE is:  2.030108 
+#The the RMSE is: 1.98311611667341
 #####################################
 ```
 
 Then we can do the O2PLS analysis with nc = 5, nx = 3, ny =3. You can also select the best parameters by looking at the cross validation results.
 ```{r}
-fit <- o2pls(X,Y,5,3,3)
+fit <- o2pls(X,Y,1,2,3)
 summary(fit)
-######### Summary of the O2PLS results #########
-### Call o2pls(X, Y, nc= 5 , nx= 3 , ny= 3 ) ###
-### Total variation 
-### X: 4900 ; Y: 4900  ###
-### Total modeled variation ### X: 0.286 ; Y: 0.304  ###
-### Joint, Orthogonal, Noise (proportions) ###
-#               X     Y
-#Joint      0.176 0.192
-#Orthogonal 0.110 0.112
-#Noise      0.714 0.696
-### Variation in X joint part predicted by Y Joint part: 0.906 
-### Variation in Y joint part predicted by X Joint part: 0.908 
-### Variation in each Latent Variable (LV) in Joint part: 
-#      LV1     LV2     LV3     LV4     LV5
-#X 181.764 179.595 191.210 152.174 157.819
-#Y 229.308 204.829 175.926 173.382 155.934
-### Variation in each Latent Variable (LV) in X Orthogonal part: 
-#      LV1     LV2     LV3
-#X 227.856 166.718 143.602
-### Variation in each Latent Variable (LV) in Y Orthogonal part: 
-#      LV1     LV2     LV3
-#Y 225.833 166.231 157.976
 
+
+############################################
 ```
 
 Extract the loadings and scores from the fit results and generated figures
@@ -127,7 +101,7 @@ plot(fit,type="loading",var="Xjoint", group=group,repel=F,rotation=TRUE)
 
 Do the OPLSDA based on the O2PLS results and calculate the VIP values
 ```{r}
-res <- oplsda(fit,group, nc=5)
+res <- oplsda(fit,group, nc=1)
 plot(res,type="score", group=group)
 vip <- vip(res)
 plot(res,type="vip", group = group, repel = FALSE,order=TRUE)