Merge branch 'master' into develop

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: mdatools
 Title: Multivariate Data Analysis for Chemometrics
-Version: 0.8.0
-Date: 2016-10-07
+Version: 0.8.1
+Date: 2016-10-30
 Author: Sergey Kucheryavskiy
 Maintainer: Sergey Kucheryavskiy <[email protected]>
 Description: Package implements projection based methods for preprocessing,

NEWS

@@ -1,3 +1,9 @@
+v.0.8.1
+=======
+* fixed a bug in PCA when explained variance was calculated incorrectly for data with excluded rows
+* fixed several issues with SIMCA (cross-validation) and SIMCAM (Cooman's plot)
+* added a chapter about SIMCA to the tutorial
+
 v.0.8.0
 =======
 * tutorial has been moved from GitBook to Bookdown and fully rewritten

NEWS.md

@@ -1,3 +1,9 @@
+v.0.8.1
+=======
+* fixed a bug in PCA when explained variance was calculated incorrectly for data with excluded rows
+* fixed several issues with SIMCA (cross-validation) and SIMCAM (Cooman's plot)
+* added a chapter about SIMCA to the tutorial
+
 v.0.8.0
 =======
 * tutorial has been moved from GitBook to Bookdown and fully rewritten

R/ldecomp.R

@@ -238,11 +238,13 @@ ldecomp.getVariances = function(Q, totvar) {
    cumresvar = colSums(Q) / totvar * 100
    cumexpvar = 100 - cumresvar
    expvar = c(cumexpvar[1], diff(cumexpvar))
-   
+
    res = list(
       expvar = expvar,
       cumexpvar = cumexpvar
    )
+
+   res
 }
 
 #' Statistical limits for Q and T2 residuals

R/pca.R

@@ -685,13 +685,17 @@ predict.pca = function(object, x, cal = FALSE, ...) {
    x = prep.autoscale(x, center = object$center, scale = object$scale)
    scores = x %*% object$loadings
    residuals = x - tcrossprod(scores, object$loadings)
-      
+
    # compute total variance
    if (length(object$exclcols) > 0){
       x = x[, -object$exclcols, drop = F]
       attrs$exclcols = object$exclcols      
    }
 
+   if (length(attrs$exclrows) > 0){
+      x = x[-object$exclrows, ,drop = F]
+   }
+
    totvar = sum(x^2)
 
    # create and return the results object

R/simca.R

@@ -324,20 +324,14 @@ simca.crossval = function(model, x, cv, center = T, scale = F) {
             Q[ind, ] = Q[ind, ] + res$Q
             T2[ind, ] = T2[ind, ] + res$T2
 
-            # compute limits
-            lim = ldecomp.getResLimits(m$eigenvals, nrow(x.cal), ncomp, model$alpha)
-            Qlim = Qlim + lim$Qlim
-            T2lim = T2lim + lim$T2lim
          }
       }  
    }
 
    Q = Q / nrep;
    T2 = T2 / nrep;
-   Qlim = Qlim / nrep;
-   T2lim = T2lim / nrep;
 
-   m = list(Qlim = Qlim, T2lim = T2lim, classname = model$classname, ncomp = model$ncomp)
+   m = list(Qlim = model$Qlim, T2lim = model$T2lim, classname = model$classname, ncomp = model$ncomp)
    r = list(Q = Q, T2 = T2, classname = model$classname)
    c.pred = simca.classify(m, r)
 

R/simcam.R

@@ -166,7 +166,6 @@ predict.simcam = function(object, x, c.ref = NULL, cv = F, ...) {
       pred.res[[i]] = predict(object$models[[i]], x, c.ref)
       c.pred[, , i] = pred.res[[i]]$c.pred[, object$models[[i]]$ncomp.selected, ]
    }
-
    dimnames(c.pred) = list(rownames(x), paste('Comp'), object$classnames)
    c.pred = mda.setattr(c.pred, attrs, 'row')
    attr(c.pred, 'name') = 'SIMCAM predictions'

R/simcamres.R

@@ -179,7 +179,7 @@ plotCooman.simcamres = function(obj, nc = c(1, 2), type = 'p', main = "Cooman's
    attrs = mda.getattr(obj$c.pred)
    res1 = obj$pred.res[[nc[1]]]
    res2 = obj$pred.res[[nc[2]]]
-   data = cbind(sqrt(res1$Q[, res1$ncomp.selected]), sqrt(res2$Q[, res2$ncomp.selected]))
+   data = cbind(res1$Q[, res1$ncomp.selected], res2$Q[, res2$ncomp.selected])
    rownames(data) = rownames(obj$c.pred)
    data = mda.setattr(data, attrs, 'row') 
    if (show.limits == T)

docs/attributes-and-factors.html

@@ -25,7 +25,7 @@
 
 <meta name="author" content="Sergey Kucheryavskiy">
 
-<meta name="date" content="2016-10-16">
+<meta name="date" content="2016-10-30">
 
 <script type="text/x-mathjax-config">
 MathJax.Hub.Config({
@@ -107,7 +107,7 @@
 <ul class="summary">
 <li class="chapter" data-level="" data-path="index.html"><a href="index.html"><i class="fa fa-check"></i>Introduction</a></li>
 <li class="chapter" data-level="" data-path="what-is-new.html"><a href="what-is-new.html"><i class="fa fa-check"></i>What is new</a><ul>
-<li class="chapter" data-level="" data-path="what-is-new.html"><a href="what-is-new.html#what-to-expect-from-future-releases"><i class="fa fa-check"></i>What to expect from future releases</a></li>
+<li class="chapter" data-level="" data-path="what-is-new.html"><a href="what-is-new.html#what-expect-from-future-releases"><i class="fa fa-check"></i>What expect from future releases</a></li>
 <li class="chapter" data-level="" data-path="what-is-new.html"><a href="what-is-new.html#last-changes-and-improvements-to-book-and-package"><i class="fa fa-check"></i>Last changes and improvements to book and package</a></li>
 </ul></li>
 <li class="chapter" data-level="" data-path="overview.html"><a href="overview.html"><i class="fa fa-check"></i>Overview</a><ul>
@@ -159,6 +159,12 @@
 </ul></li>
 <li class="chapter" data-level="" data-path="variable-selection.html"><a href="variable-selection.html"><i class="fa fa-check"></i>Variable selection</a></li>
 </ul></li>
+<li class="chapter" data-level="" data-path="simca-classification.html"><a href="simca-classification.html"><i class="fa fa-check"></i>SIMCA classification</a><ul>
+<li class="chapter" data-level="" data-path="calibration-and-validation.html"><a href="calibration-and-validation.html"><i class="fa fa-check"></i>Calibration and validation</a><ul>
+<li class="chapter" data-level="" data-path="calibration-and-validation.html"><a href="calibration-and-validation.html#predictions-and-validation-with-a-test-set"><i class="fa fa-check"></i>Predictions and validation with a test set</a></li>
+</ul></li>
+<li class="chapter" data-level="" data-path="multiclass-classification.html"><a href="multiclass-classification.html"><i class="fa fa-check"></i>Multiclass classification</a></li>
+</ul></li>
 </ul>
 
       </nav>