Differences

This shows you the differences between two versions of the page.

--- principal_component_analysis [2019/11/14 21:00] – hkimscil
+++ principal_component_analysis [2019/11/16 06:06] (current) – [e.g. saq] hkimscil
@@ Line 1: / Line 1: @@
 ====== PCA ======
+[[https://www.theanalysisfactor.com/the-fundamental-difference-between-principal-component-analysis-and-factor-analysis/|Difference between PCA and FA]]
+  * Both are data reduction techniques — they allow you to capture the variance in variables in a smaller set.
+  * Both are usually run in stat software using the same procedure, and the output looks pretty much the same.
+  * The steps you take to run them are the same-extraction, interpretation, rotation, choosing the number of factors or components.
+  * Despite all these similarities, there is a fundamental difference between them: PCA is a linear combination of variables; Factor Analysis is a measurement model of a latent variable.
+====== Some useful lectures ======
 {{youtube>FgakZw6K1QQ}}
 <WRAP clear />
@@ Line 153: / Line 160: @@
   ggtitle("eigen on cov(t(data.matrix))")
 </code>
+====== e.g. saq ======
+SPSS Anxiety Questionnaire
+{{:r:saq8.csv}}
+<code>
+# saq <- read.csv("http://commres.net/wiki/_media/r/saq.csv", header = T)
+saq8 <- read.csv("http://commres.net/wiki/_media/r/saq8.csv", header = T)
+head(saq8)
+saq8 <- saq8[c(-1)]
+</code>
+<code>
+> round(cor(saq8),3)
+              stat_cry afraid_spss sd_excite nmare_pearson du_stat lexp_comp comp_hate good_math
+stat_cry         1.000      -0.099    -0.337         0.436   0.402     0.217     0.305     0.331
+afraid_spss     -0.099       1.000     0.318        -0.112  -0.119    -0.074    -0.159    -0.050s
+sd_excite       -0.337       0.318     1.000        -0.380  -0.310    -0.227    -0.382    -0.259
+nmare_pearson    0.436      -0.112    -0.380         1.000   0.401     0.278     0.409     0.349
+du_stat          0.402      -0.119    -0.310         0.401   1.000     0.257     0.339     0.269
+lexp_comp        0.217      -0.074    -0.227         0.278   0.257     1.000     0.514     0.223
+comp_hate        0.305      -0.159    -0.382         0.409   0.339     0.514     1.000     0.297
+good_math        0.331      -0.050    -0.259         0.349   0.269     0.223     0.297     1.000
+>
+</code>
+<code>
+install.packages("Hmisc")
+library(Hmisc)
+saq8.rcorr <- rcorr(as.matrix(saq8))
+print(saq8.rcorr$r, digits = 3)
+</code>