Zusammenfassung
Liegen von Beobachtungsobjekten Werte mehrerer Variablen vor, kann sich die Datenanalyse nicht nur auf jede Variable einzeln, sondern auch auf die gemeinsame Verteilung der Variablen beziehen. Solche Fragestellungen sind mit multivariaten Verfahren zu bearbeiten (Backhaus, Erichson, Plinke, & Weiber, 2015a Backhaus, Erichson, & Weiber, 2015b; Mardia, Kent, & Bibby, 1980), deren Anwendung in R Zelterman (2015) vertiefend behandelt. Abschn. 14.6.8, 14.7 und 15.3 thematisieren Möglichkeiten, multivariate Daten in Diagrammen zu veranschaulichen.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Die hier vorgestellten Rechenwege setzen die mathematischen Formeln direkt um. Tatsächlich gibt es häufig numerisch effizientere und stabilere Möglichkeiten, um dieselben Ergebnisse zu erhalten. So entspricht die Implementierung von R-eigenen Funktionen auch meist nicht den hier vorgestellten Rechnungen (Bates, 2004). Siehe Abschn. 17.3.3 für Wege, die Effizienz der Berechnungen zu steigern. Das Paket Matrix (Bates & Maechler, 2016) enthält fortgeschrittene Methoden der Matrix-Algebra – etwa zu schwachbesetzten Matrizen, die u. a. als Designmatrizen linearer Modelle auftauchen (Abschn. 12.9.1).
- 2.
Bei nicht quadratischen (p × q)-Matrizen \(\mathbf{X}\) sind dies die Elemente x 11, …, x pp (für p < q) bzw. x 11, …, x qq (für p > q).
- 3.
Eine Dezimalzahl wird dabei tranchiert. Eine (1 × 1)-Diagonalmatrix \(\mathbf{X}\) kann mit einem weiteren Argument als diag(x, nrow=1) erzeugt werden.
- 4.
Mit dem Operator 〈Matrix〉 %^% 〈Zahl〉 aus dem Paket expm (Goulet et al., 2015) ist das Exponenzieren von quadratischen Matrizen möglich, mit logm() aus demselben Paket das Logarithmieren.
- 5.
Sie ist zudem numerisch effizienter als die Verwendung von t(X) %*% Y. Die Benennung erscheint unglücklich, da Verwechslungen mit dem Vektor-Kreuzprodukt naheliegen, das man stattdessen mit cross() aus dem Paket pracma (Borchers, 2016b) erhält.
- 6.
Im Anwendungsfall würde man stattdessen auf scale(〈Matrix〉, center=TRUE, scale=FALSE) zurückgreifen, um eine Matrix spaltenweise zu zentrieren.
- 7.
Für die Pseudoinverse \(\mathbf{A}^{+}\) einer nicht invertierbaren Matrix \(\mathbf{A}\) vgl. ginv() aus dem MASS Paket. Für solche Matrizen ermittelt Null() aus demselben Paket eine Basis des Kerns von \(\mathbf{A}^{\top }\) (null space).
- 8.
Jede Transformation der Form \(\mathbf{G}\mathbf{S}^{-\frac{1} {2} }(\mathbf{x} -\overline{\mathbf{x}})\) mit \(\mathbf{G}\) als Orthogonalmatrix (\(\mathbf{G}^{\top } = \mathbf{G}^{-1}\)) würde ebenfalls eine multivariate z-Transformation liefern.
- 9.
Es sei \(\mathbf{a}_{i\cdot }\) die i-te Zeile und \(\mathbf{a}_{\cdot j}\) die j-te Spalte von \(\mathbf{A}\). Dann gilt \(\mbox{ tr}(\mathbf{A}^{\top }\mathbf{A}) =\sum \nolimits_{j}\mathbf{a}_{\cdot j}^{\top }\mathbf{a}_{\cdot j} =\sum \nolimits_{j}\sum \nolimits_{i}a_{ij}^{2} =\sum \nolimits_{i}\sum \nolimits_{j}a_{ij}^{2} =\sum \nolimits_{i}\mathbf{a}_{i\cdot }^{\top }\mathbf{a}_{i\cdot } = \mbox{ tr}(\mathbf{A}\mathbf{A}^{\top })\).
- 10.
\(\det (\mathbf{A}) \cdot \det (\mathbf{A}^{-1}) =\det (\mathbf{A}\mathbf{A}^{-1}) =\det (\mathbf{I}) =\prod \nolimits_{ i=1}^{p}1 = 1^{p} = 1\).
- 11.
Eigenwerte werden entsprechend ihrer algebraischen Vielfachheit ggf. mehrfach aufgeführt. Auch Matrizen mit komplexen Eigenwerten sind zugelassen. Da in der Statistik vor allem Eigenwerte von Kovarianzmatrizen interessant sind, konzentriert sich die Darstellung hier auf den Fall reeller symmetrischer Matrizen. Ihre Eigenwerte sind alle reell, zudem stimmen algebraische und geometrische Vielfachheiten überein.
- 12.
\(\mathbf{A}\mathbf{A} = \mathbf{G}\mathbf{D}^{\frac{1} {2} }\mathbf{G}^{\top }\mathbf{G}\mathbf{D}^{\frac{1} {2} }\mathbf{G}^{\top } = \mathbf{G}\mathbf{D}^{\frac{1} {2} }\mathbf{D}^{\frac{1} {2} }\mathbf{G}^{\top } = \mathbf{G}\mathbf{D}\mathbf{G}^{\top } = \mathbf{X}\). Siehe auch sqrtm() aus dem Paket expm.
- 13.
\(\mathbf{N}\mathbf{N}^{\top } = \mathbf{G}\mathbf{D}^{\frac{1} {2} }(\mathbf{G}\mathbf{D}^{\frac{1} {2} })^{\top } = \mathbf{G}\mathbf{D}^{\frac{1} {2} }(\mathbf{D}^{\frac{1} {2} })^{\top }\mathbf{G}^{\top } = \mathbf{G}\mathbf{D}\mathbf{G}^{\top } = \mathbf{X}\).
- 14.
(A ⊤ A)−1 A ⊤ ist die Pseudoinverse A + von A.
- 15.
Im Kontext linearer Modelle ist P die Hat-Matrix (Abschn. 6.3.1).
- 16.
Ist v aus V, lässt sich v = Cy schreiben, wobei y der Koordinatenvektor von v bzgl. einer Orthogonalbasis C von V ist. Damit folgt \(\mathbf{P}\mathbf{v} = \mathbf{C}\mathbf{C}^{\top }\mathbf{v} = \mathbf{C}\mathbf{C}^{\top }\mathbf{C}\mathbf{y} = \mathbf{C}\mathbf{y} = \mathbf{v}\).
- 17.
Zunächst gilt P 2 = (CC ⊤)(CC ⊤) = CC ⊤ = P. Weiter gilt P ⊤ = (CC ⊤)⊤ = CC ⊤ = P.
- 18.
tr(P) = tr(CC ⊤) = tr(C ⊤ C) = tr(I) = p.
- 19.
\(\mathbf{P}_{1}\mathbf{P}_{2} = (\mathbf{P}_{1}\mathbf{P}_{2})^{\top } = \mathbf{P}_{2}^{\top }\mathbf{P}_{1}^{\top } = \mathbf{P}_{2}\mathbf{P}_{1}\).
- 20.
Ist w = (I −P)x aus \(V ^{\perp }\) und v aus V, muss w ⊤ v = 0 gelten. v lässt sich als v = Cy schreiben, wobei y der Koordinatenvektor von v bzgl. einer Orthogonalbasis C von V ist. Nun gilt \(\mathbf{w}^{\top }\mathbf{v} = ((\mathbf{I}-\mathbf{P})\mathbf{x})^{\top }\mathbf{C}\mathbf{y} = \mathbf{x}^{\top }(\mathbf{I}-\mathbf{P})\mathbf{C}\mathbf{y} = (\mathbf{x}^{\top }-\mathbf{x}^{\top }\mathbf{C}\mathbf{C}^{\top })\mathbf{C}\mathbf{y} = \mathbf{x}^{\top }\mathbf{C}\mathbf{y}-\mathbf{x}^{\top }\mathbf{C}\mathbf{C}^{\top }\mathbf{C}\mathbf{y} = \mathbf{x}^{\top }\mathbf{C}\mathbf{y}-\mathbf{x}^{\top }\mathbf{C}\mathbf{y} = \mathbf{0}\).
- 21.
(I −P)v = Iv −Pv = v −v = 0, da v durch P unverändert bleibt.
- 22.
\(\mathbf{P}_{1} = \mathbf{1}_{n}(\mathbf{1}_{n}^{\top }\mathbf{1}_{n})^{-1}\mathbf{1}_{n}^{\top } = \frac{1} {n}\,\mathbf{1}_{n}\mathbf{1}_{n}^{\top } = \frac{1} {n}\,\mathbf{1}_{n\times n}\).
- 23.
Hier sei voller Spaltenrang von X vorausgesetzt. Dann ist \(\mathbf{X}^{+} = (\mathbf{X}^{\top }\mathbf{X})^{-1}\mathbf{X}^{\top } = ((\mathbf{Q}\mathbf{R})^{\top }\mathbf{Q}\mathbf{R})^{-1}(\mathbf{Q}\mathbf{R})^{\top } = (\mathbf{R}^{\top }\mathbf{Q}^{\top }\mathbf{Q}\mathbf{R})^{-1}\mathbf{R}^{\top }\mathbf{Q}^{\top } = (\mathbf{R}^{\top }\mathbf{R})^{-1}\mathbf{R}^{\top }\mathbf{Q}^{\top } = \mathbf{R}^{-1}\mathbf{R}^{t-1}\mathbf{R}^{\top }\mathbf{Q}^{\top } = \mathbf{R}^{-1}\mathbf{Q}^{\top }\).
- 24.
\(\mathbf{X}\mathbf{X}^{+} = \mathbf{Q}\mathbf{R}\mathbf{R}^{-1}\mathbf{Q}^{\top } = \mathbf{Q}\mathbf{Q}^{\top }\).
- 25.
Es sei \(\overline{\mathbf{x}}\) das Zentroid der spaltenweise aus den Variablen zusammengestellten Datenmatrix X, x ein Datenvektor und G die Matrix der spaltenweise zusammengestellten normierten Eigenvektoren der Kovarianzmatrix S von X (Abschn. 12.1.5). Mit dem Spektralsatz ist G eine Orthogonalmatrix (\(\mathbf{G}^{\top } = \mathbf{G}^{-1}\), s. Abschn. 12.1.6). Dann berechnet sich der zugehörige Vektor y der Hauptkomponenten als \(\mathbf{y} = \mathbf{G}^{\top }(\mathbf{x} -\overline{\mathbf{x}})\).
- 26.
Genauer gesagt ist ihre Kovarianz 0 – da ihre Varianz auch 0 sein kann, ist die Korrelation nicht immer definiert. S lässt sich mit \(\mathbf{S} = \mathbf{G}\mathbf{D}\mathbf{G}^{\top }\) diagonalisieren (Abschn. 12.1.6). Dabei ist D die zu G gehörende, aus den Eigenwerten von S gebildete Diagonalmatrix. Damit gilt \(V (\mathbf{y}) = V (\mathbf{G}^{\top }(\mathbf{x}-\overline{\mathbf{x}})) = V (\mathbf{G}^{\top }\mathbf{x}-\mathbf{G}^{\top }\overline{\mathbf{x}}) = V (\mathbf{G}^{\top }\mathbf{x}) = \mathbf{G}^{\top }V (\mathbf{x})\mathbf{G} = \mathbf{G}^{\top }\mathbf{S}\mathbf{G} = \mathbf{G}^{\top }\mathbf{G}\mathbf{D}\mathbf{G}^{\top }\mathbf{G} = \mathbf{D}\). Die Varianzen der Hauptkomponenten sind also gleich den Eigenwerten der Kovarianzmatrix der Daten, die Kovarianzen sind 0.
- 27.
Es seien x 1 und x 2 zwei Datenvektoren mit zugehörigen Hauptkomponenten-Vektoren y 1 und y 2. Dann gilt \(\|\mathbf{x}_{2}-\mathbf{x}_{1}\|^{2} = (\mathbf{x}_{2}-\mathbf{x}_{1})^{\top }(\mathbf{x}_{2}-\mathbf{x}_{1}) = (\mathbf{x}_{2}-\mathbf{x}_{1})^{\top }\mathbf{G}\mathbf{G}^{\top }(\mathbf{x}_{2}-\mathbf{x}_{1}) = (\mathbf{G}^{\top }(\mathbf{x}_{2}-\mathbf{x}_{1}))^{\top }\mathbf{G}^{\top }(\mathbf{x}_{2}-\mathbf{x}_{1}) =\| \mathbf{G}^{\top }(\mathbf{x}_{2}-\mathbf{x}_{1})\|^{2} =\| \mathbf{G}^{\top }\mathbf{x}_{2}-\mathbf{G}^{\top }\mathbf{x}_{1}\|^{2} =\| \mathbf{G}^{\top }(\mathbf{x}_{2}-\overline{\mathbf{x}})-\mathbf{G}^{\top }(\mathbf{x}_{1}-\overline{\mathbf{x}})\|^{2} =\| \mathbf{y}_{2}-\mathbf{y}_{1}\|^{2}\).
- 28.
Für die robuste Hauptkomponentenanalyse vgl. das Paket pcaPP (Filzmoser, Fritz, & Kalcher, 2016).
- 29.
Zunächst gilt \(\mathbf{H} = \mathbf{Y}\mathbf{D}^{-\frac{1} {2} }\), wobei \(\mathbf{Y} =\dot{ \mathbf{X}}\mathbf{G} = \mathbf{Z}\mathbf{X}\mathbf{G}\) die Matrix der (zentrierten) Hauptkomponenten und Z die Zentriermatrix zu X ist (Abschn. 12.1.1). Wegen \(\mathbf{G}^{\top } = \mathbf{G}^{-1}\) folgt \(\mathbf{H}\mathbf{B}^{\top } + \overline{\mathbf{x}} = \mathbf{Y}\mathbf{D}^{-\frac{1} {2} }(\mathbf{G}\mathbf{D}^{\frac{1} {2} })^{\top } + \overline{\mathbf{x}} = \mathbf{Z}\mathbf{X}\mathbf{G}\mathbf{D}^{-\frac{1} {2} }\mathbf{D}^{\frac{1} {2} }\mathbf{G}^{\top } + \overline{\mathbf{x}} =\dot{ \mathbf{X}}\mathbf{G}\mathbf{G}^{\top } + \overline{\mathbf{x}} = \mathbf{X}\).
- 30.
\(\mathbf{B}\mathbf{B}^{\top } = \mathbf{G}\mathbf{D}^{\frac{1} {2} }(\mathbf{G}\mathbf{D}^{\frac{1} {2} })^{\top } = \mathbf{G}\mathbf{D}^{\frac{1} {2} }\mathbf{D}^{\frac{1} {2} }\mathbf{G}^{\top } = \mathbf{G}\mathbf{D}\mathbf{G}^{\top } = \mathbf{S}\).
- 31.
Weiterhin basiert prcomp() intern auf der Singulärwertzerlegung mit svd(), princomp() hingegen auf der Berechnung der Eigenwerte mit eigen() (Abschn. 12.1.6). Die Singulärwertzerlegung gilt als numerisch stabiler bei schlecht konditionierten Matrizen i. S. der Kondition κ (Abschn. 12.1.5) – prcomp() sollte also vorgezogen werden.
- 32.
Für weitere Verfahren, die die Beziehungen latenter und beobachtbarer Variablen modellieren, vgl. den Abschnitt Psychometric Models der CRAN Task Views (Mair, 2016). Lineare Strukturgleichungsmodelle werden durch die Pakete sem (Fox, Nie, & Byrnes, 2016), OpenMx (Boker et al., 2011) und lavaan (Rosseel, 2012) unterstützt.
- 33.
In diesem Sinne ist die Faktorenanalyse das Gegenteil der Hauptkomponentenanalyse, in der die Hauptkomponenten Linearkombinationen der beobachtbaren Variablen sind.
- 34.
Für n Personen seien die Werte auf den beobachtbaren Variablen zeilenweise in einer (n × p)-Matrix X zusammengefasst, analog die Faktorwerte in einer (n × q)-Matrix F und die Fehler in einer (n × p)-Matrix E. Dann lautet das Modell \(\mathbf{X} = \mathbf{F}\boldsymbol{\Lambda }^{\top } + \mathbf{E}\).
- 35.
Die konfirmatorische Faktorenanalyse, bei der theoretische Erwägungen ein bestimmtes, auf Konsistenz mit den Daten zu testendes \(\boldsymbol{\Lambda }\) vorgeben, ist mit Hilfe linearer Strukturgleichungsmodelle durchzuführen (Fußnote 32).
- 36.
Weitere Rotationsarten, etwa für das Modell korrelierter Faktoren, stellt das Paket GPArotation (Bernaards & Jennrich, 2005) zur Verfügung. Es enthält Funktionen, deren Namen an das Argument rotation übergeben werden können, z. B. "oblimin" für eine schiefwinklige Rotation. Für eine vollständige Liste vgl. ?rotations, nachdem das Paket installiert und geladen wurde.
- 37.
Setzt man rotation="none", ist dies ist bei der durch factanal() verwendeten Methode, um ein \(\hat{\boldsymbol{\Lambda }}\) zu erzeugen, gleichzeitig der zugehörige Eigenwert der geschätzten reduzierten Korrelationsmatrix \(\hat{\mathbf{K}}_{\hat{\mathbf{x}}} =\hat{\boldsymbol{ \Lambda }}\hat{\boldsymbol{\Lambda }}^{\top }\). Bei der Methode handelt es sich um die iterative Maximum-Likelihood Kommunalitätenschätzung. Bei Rotation oder anderen Schätzmethoden gilt diese Gleichheit dagegen nicht.
- 38.
Die Summe der Eigenwerte einer Matrix ist gleich deren Spur (Abschn. 12.1.5) – im Fall der Korrelationsmatrix K x also gleich der Anzahl der Variablen p, da in der Diagonale überall 1 steht. Der Mittelwert der Eigenwerte ist damit \(\frac{1} {p}p = 1\).
- 39.
Weitere Verfahren, die der Klärung der geeigneten Anzahl von Faktoren dienen sollen, sind die Parallelanalyse mit fa.parallel() aus dem Paket psych sowie das very simple structure Verfahren mit VSS() aus demselben Paket.
- 40.
Die nichtmetrische multidimensionale Skalierung wird durch monoMDS() aus dem vegan Paket bereit gestellt.
- 41.
Bei der multivariaten Formulierung des Modells wird intern aufgrund der generischen anova() Funktion automatisch anova.mlm() verwendet, ohne dass dies explizit angegeben werden muss (Abschn. 17.2.6).
- 42.
Auch bei der multivariaten zweifaktoriellen Varianzanalyse ist im Fall ungleicher Zellbesetzungen zu beachten, dass R in der Voreinstellung Quadratsummen vom Typ I berechnet (Abschn. 7.5.2 und 12.9.6). Manova() aus dem car Paket erlaubt es, analog zur Verwendung von Anova() (Abschn. 7.4.3), Quadratsummen vom Typ II und III zu berechnen.
- 43.
Lässt man auch quadratische Funktionen der ursprünglichen Variablen zu, ergibt sich die quadratische Diskriminanzanalyse. Sie wird mit qda() aus dem MASS Paket berechnet.
- 44.
Für weitere Klassifikationsverfahren wie Varianten der Clusteranalyse, CART-Modelle oder support vector machines vgl. die Abschnitte Cluster Analysis (Leisch & Gruen, 2016), Multivariate Statistics (Hewson, 2015) und Machine Learning & Statistical Learning (Hothorn, 2016) der CRAN Task Views. Die logistische und multinomiale Regression (Abschn. 8.1 und 8.3) lassen sich ebenfalls zur Klassifikation verwenden und besitzen weniger Verteilungsvoraussetzungen als die Diskriminanzanalyse.
- 45.
- 46.
Die x j sind feste Realisierungen eines Zufallsvektors, also stochastische Prädiktoren. Sie enthalten damit nicht alle möglichen Prädiktorwerte, sondern nur jeweils n viele. Man könnte daher auch vom Vektor E(y | X) der auf eine konkrete Designmatrix X bedingten Erwartungswerte von y sprechen, worauf hier aber verzichtet wird. Die x j müssen fehlerfrei sein, bei den x ij muss es sich also um die wahren Prädiktorwerte handeln. Ohne diese Annahme kommen lineare Strukturgleichungsmodelle zur Auswertung in Betracht (Abschn. 12.3, Fußnote 32).
- 47.
Anders als in der Regression werden die Werte der Indikatorvariablen hier durch die Zuordnung von Beobachtungen zu Gruppen systematisch hergestellt, sind also keine stochastischen Prädiktoren.
- 48.
Eine der folgenden Zusammenfassung ähnliche Exposition enthält die Vignette des Pakets codingMatrices (Venables, 2016). Dabei entsprechen sich folgende Bezeichnungen: Die Inzidenzmatrix \(\mathbf{X}_{p}^{\star }\) hier ist dort X. Die reduzierte Designmatrix X hier ist dort \(\widetilde{\mathbf{X}}\), die Codiermatrix C hier ist dort \(\mathbf{B}_{\star }\), die Matrix [1 | C] hier ist dort \(\mathbf{B} = [\mathbf{1}\vert \mathbf{B}_{\star }] = \mathbf{C}^{-1}\), der Parametervektor \(\boldsymbol{\beta }_{p-1}\) hier ist dort \(\boldsymbol{\beta }_{\star }\).
- 49.
Alternativ kann auch der Parameter β 0 = 0 gesetzt werden, womit \(\mathbf{X}^{\star } = \mathbf{X}_{p}^{\star }\) ist. Diese Möglichkeit zur Parametrisierung soll hier nicht weiter verfolgt werden, um das Modell wie jenes der Regression formulieren zu können (Fußnote 53).
- 50.
- 51.
Denn dann gilt \(\mathbf{v}^{\top }\boldsymbol{\beta }_{p}^{\star } = \mathbf{v}^{\top }\mathbf{C}\boldsymbol{\beta }_{p-1} = \mathbf{0}^{\top }\boldsymbol{\beta }_{p-1} = 0\). v steht senkrecht auf den Spalten von C, ist also eine Basis des orthogonalen Komplements des von den Spalten von C aufgespannten Unterraums. Mit anderen Worten ist v wegen \(\mathbf{0} = (\mathbf{v}^{\top }\mathbf{C})^{\top } = \mathbf{C}^{\top }\mathbf{v}\) eine Basis des Kerns von \(\mathbf{C}^{\top }\).
- 52.
Die in einem linearen Modell mit kategorialen Variablen von R weggelassene Gruppe ist die erste Stufe von levels(〈Faktor〉).
- 53.
Die in Fußnote 49 erwähnte Möglichkeit der Parametrisierung führt zum cell means Modell, bei dem \(\mathbf{X} = \mathbf{X}^{\star } = \mathbf{X}_{p}^{\star }\) gilt und die Parameter \(\beta _{j}^{\star }\) direkt die Bedeutung der Gruppenerwartungswerte μ j erhalten.
- 54.
Ein andere Wahl für C unter der Nebenbedingung v = 1 ist die Helmert-Codierung mit paarweise orthogonalen Spalten von [1 | C] (vgl. ?contr.helmert). Die Parameter haben dann jedoch eine andere Bedeutung.
- 55.
Alternativ ließe sich \(\mu =\sum \nolimits_{j}\frac{n_{j}} {n} \,\mu _{j}\) als mit den anteiligen Zellbesetzungen \(\frac{n_{j}} {n}\) gewichtetes Mittel der μ j definieren. Die zugehörige Nebenbedingung für die β j ⋆ i. S. der α j lautet dann \(\sum \nolimits_{j}\frac{n_{j}} {n} \,\beta _{j}^{\star } = 0\), d. h. \(\mathbf{v} = (\frac{n_{1}} {n},\ldots, \frac{n_{p}} {n} )^{\top }\). Diese Parametrisierung lässt sich mit der gewichteten Effektcodierung umsetzen, die zunächst der ungewichteten gleicht. In der letzten Zeile der Matrix C erhalten hier jedoch Mitglieder der Gruppe p für die X j nicht den Wert − 1, sondern \(-\frac{n_{j}} {n_{p}}\).
- 56.
Hierbei ist die Reihenfolge relevant, mit denen man die Spalten von X 1×2 ⋆ und entsprechend die Parameter in \(\boldsymbol{\beta }_{1\times 2}^{\star }\) ordnet, die zu den Kombinationen jk der Faktorstufen der UV 1 und UV 2 gehören: Variiert (wie hier) der erste Index j schnell und der zweite Index k langsam (\(\boldsymbol{\beta }_{1\times 2}^{\star } = (\beta _{11}^{\star },\beta _{21}^{\star },\beta _{31}^{\star },\;\beta _{12}^{\star },\beta _{22}^{\star },\beta _{32}^{\star })^{\top }\)), gilt C 1×2 = C 2 ⊗C 1. Dies ist die Voreinstellung in R, die sich etwa an der Ausgabe von interaction(〈UV1〉, 〈UV2〉) zeigt (Abschn. 2.6.2). Variiert dagegen j langsam und k schnell (\(\boldsymbol{\beta }_{1\times 2}^{\star } = (\beta _{11}^{\star },\beta _{12}^{\star },\;\beta _{21}^{\star },\beta _{22}^{\star },\;\beta _{31}^{\star },\beta _{32}^{\star })^{\top }\)), ist \(\mathbf{C}_{1\times 2} = \mathbf{C}_{1} \otimes \mathbf{C}_{2}\) zu setzen.
- 57.
Hierbei seien α j = μ j. −μ die Effektgrößen des Haupteffekts der ersten UV, β k = μ . k −μ die des Haupteffekts der zweiten UV und (α β) jk = μ jk − (μ +α j +β k ) = μ jk −μ j. −μ . k +μ die der Interaktion. Dabei seien \(\mu _{j.} = \frac{1} {q}\sum \nolimits_{k}\mu _{jk}\), \(\mu _{.k} = \frac{1} {p}\sum \nolimits_{j}\mu _{jk}\) und \(\mu = \frac{1} {p\cdot q}\sum \nolimits_{j}\sum \nolimits_{k}\mu _{jk}\) ungewichtete mittlere Erwartungswerte. Liegen gleiche, oder zumindest proportional ungleiche Zellbesetzungen vor (\( \frac{n_{jk}} {n_{jk^{\prime}}} = \frac{n_{j^{\prime}k}} {n_{j^{\prime}k^{\prime}}} \) sowie \( \frac{n_{jk}} {n_{j^{\prime}k}} = \frac{n_{jk^{\prime}}} {n_{j^{\prime} k^{\prime}}} \) für alle j, j′, k, k′), lässt sich die Parametrisierung mit gewichteten mittleren Erwartungswerten durch die gewichtete Effektcodierung umsetzen (Fußnote 55).
- 58.
Zunächst ist \(E(\mathbf{y}) = E(\mathbf{X}\boldsymbol{\beta }+\boldsymbol{\epsilon }) = \mathbf{X}\boldsymbol{\beta } + E(\boldsymbol{\epsilon }) = \mathbf{X}\boldsymbol{\beta }\). Weiter gilt \(V (\mathbf{y}) = V (\mathbf{X}\boldsymbol{\beta }+\boldsymbol{\epsilon }) = V (\boldsymbol{\epsilon }) =\sigma ^{2}\mathbf{I}\).
- 59.
In der Varianzanalyse werden die reduzierten Parameter geschätzt. Für die Beziehung zwischen geschätzten ursprünglichen Parametern und geschätzten reduzierten Parametern gilt \(\hat{\boldsymbol{\beta }}^{\star } = [\mathbf{1}\vert \mathbf{C}]\,\hat{\boldsymbol{\beta }}\). Auf Basis eines mit aov() oder lm() angepassten linearen Modells erhält man \(\hat{\boldsymbol{\beta }}\) mit coef() und \(\hat{\boldsymbol{\beta }}^{\star }\) mit dummy.coef().
- 60.
Zunächst ist \(E(\hat{\boldsymbol{\beta }}) = E(\mathbf{X}^{+}\mathbf{y}) = \mathbf{X}^{+}E(\mathbf{y}) = (\mathbf{X}^{\top }\mathbf{X})^{-1}\mathbf{X}^{\top }\mathbf{X}\boldsymbol{\beta } =\boldsymbol{\beta }\). Weiter gilt \(V (\hat{\boldsymbol{\beta }}) = V (\mathbf{X}^{+}\mathbf{y}) = \mathbf{X}^{+}V (\mathbf{y})(\mathbf{X}^{+})^{\top } = (\mathbf{X}^{\top }\mathbf{X})^{-1}\mathbf{X}^{\top }\sigma ^{2}\mathbf{I}\mathbf{X}(\mathbf{X}^{\top }\mathbf{X})^{-1} =\sigma ^{2}(\mathbf{X}^{\top }\mathbf{X})^{-1}\).
- 61.
Zunächst ist \(E(\mathbf{e}) = E((\mathbf{I} -\mathbf{P})\mathbf{y}) = E(\mathbf{y}) - E(\mathbf{P}\mathbf{y}) = \mathbf{X}\boldsymbol{\beta } -\mathbf{P}\mathbf{X}\boldsymbol{\beta }\). Da die Spalten von X in V liegen, bleiben sie durch P unverändert, es folgt also \(E(\mathbf{e}) = \mathbf{X}\boldsymbol{\beta } -\mathbf{P}\mathbf{X}\boldsymbol{\beta } = \mathbf{0}\). Weiter gilt V (e) = V ((I −P)y) = (I −P)V (y)(I −P)⊤ = σ 2(I −P)(I −P)⊤. Als orthogonale Projektion ist I −P symmetrisch und idempotent, es gilt also V (e) = σ 2(I −P)(I −P)⊤ = σ 2(I −P).
- 62.
\(SS_{e} =\sum \nolimits_{i}e_{i}^{2} =\| \mathbf{e}\|^{2} = \mathbf{e}^{\top }\mathbf{e} = (\mathbf{y}-\hat{\mathbf{y}})^{\top }(\mathbf{y}-\hat{\mathbf{y}}) = \mathbf{y}^{\top }\mathbf{y}-\mathbf{y}^{\top }\hat{\mathbf{y}}-\hat{\mathbf{y}}^{\top }\mathbf{y}+\hat{\mathbf{y}}^{\top }\hat{\mathbf{y}} = \mathbf{y}^{\top }\mathbf{y}-\mathbf{y}^{\top }\mathbf{P}\mathbf{y}-(\mathbf{P}\mathbf{y})^{\top }\mathbf{y}+(\mathbf{P}\mathbf{y})^{\top }(\mathbf{P}\mathbf{y}) = \mathbf{y}^{\top }\mathbf{y}-\mathbf{y}^{\top }\mathbf{P}\mathbf{y}-\mathbf{y}^{\top }\mathbf{P}\mathbf{y}+\mathbf{y}^{\top }\mathbf{P}\mathbf{P}\mathbf{y} = \mathbf{y}^{\top }\mathbf{y}-\mathbf{y}^{\top }\mathbf{P}\mathbf{y} = \mathbf{y}^{\top }(\mathbf{I}-\mathbf{P})\mathbf{y}\).
- 63.
Zunächst ist \(E(\hat{\psi }) = E(\mathbf{c}^{\top }\hat{\boldsymbol{\beta }}) = \mathbf{c}^{\top }E(\hat{\boldsymbol{\beta }}) = \mathbf{c}^{\top }\boldsymbol{\beta } =\psi\). Weiter gilt \(V (\hat{\psi }) = V (\mathbf{c}^{\top }\hat{\boldsymbol{\beta }}) = \mathbf{c}^{\top }V (\hat{\boldsymbol{\beta }})\mathbf{c} = \mathbf{c}^{\top }\sigma ^{2}(\mathbf{X}^{\top }\mathbf{X})^{-1}\mathbf{c}\). Bei \(\hat{\psi }\) handelt es sich um einen Gauß-Markoff-Schätzer, also den linearen erwartungstreuen Schätzer mit der geringsten Varianz.
- 64.
Zunächst ist X ⊤ a = X ⊤ X(X ⊤ X)−1 c = c, also c ⊤ = a ⊤ X. Damit gilt \(\mathbf{a}^{\top }\mathbf{y} = \mathbf{c}^{\top }(\mathbf{X}^{\top }\mathbf{X})^{-1}\mathbf{X}^{\top }\mathbf{y} = \mathbf{a}^{\top }\mathbf{X}(\mathbf{X}^{\top }\mathbf{X})^{-1}\mathbf{X}^{\top }\mathbf{y} = \mathbf{a}^{\top }\mathbf{P}\mathbf{y} = \mathbf{a}^{\top }\hat{\mathbf{y}}\).
- 65.
\(\|\mathbf{a}\|^{2} = \mathbf{a}^{\top }\mathbf{a} = (\mathbf{X}(\mathbf{X}^{\top }\mathbf{X})^{-1}\mathbf{c})^{\top }\mathbf{X}(\mathbf{X}^{\top }\mathbf{X})^{-1}\mathbf{c} = \mathbf{c}^{\top }(\mathbf{X}^{\top }\mathbf{X})^{-1}\mathbf{X}^{\top }\mathbf{X}(\mathbf{X}^{\top }\mathbf{X})^{-1}\mathbf{c} = \mathbf{c}^{\top }(\mathbf{X}^{\top }\mathbf{X})^{-1}\mathbf{c}\).
- 66.
\(\mathbf{A}\mathbf{X}_{u} = (\mathbf{X}_{u}(\mathbf{X}_{u}^{\top }\mathbf{X}_{u})^{-1}\mathbf{L}^{\top })^{\top }\mathbf{X}_{u} = \mathbf{L}(\mathbf{X}_{u}^{\top }\mathbf{X}_{u})^{-1}\mathbf{X}_{u}^{\top }\mathbf{X}_{u} = \mathbf{L}\).
- 67.
\(\hat{\boldsymbol{\psi }}= \mathbf{L}\hat{\boldsymbol{\beta }} = \mathbf{L}(\mathbf{X}_{u}^{\top }\mathbf{X}_{u})^{-1}\mathbf{X}_{u}^{\top }\mathbf{y} = \mathbf{A}\mathbf{X}_{u}(\mathbf{X}_{u}^{\top }\mathbf{X}_{u})^{-1}\mathbf{X}_{u}^{\top }\mathbf{y} = \mathbf{A}\hat{\mathbf{y}}\).
- 68.
Zunächst gilt \(E(\mathbf{L}\hat{\boldsymbol{\beta }}) = \mathbf{L}E(\hat{\boldsymbol{\beta }}) = \mathbf{L}\boldsymbol{\beta } =\boldsymbol{\psi }\). Weiter ist \(V (\hat{\boldsymbol{\psi }}) = V (\mathbf{L}\hat{\boldsymbol{\beta }}) = \mathbf{L}V (\hat{\boldsymbol{\beta }})\mathbf{L}^{\top } =\sigma ^{2}\mathbf{L}(\mathbf{X}_{u}^{\top }\mathbf{X}_{u})^{-1}\mathbf{L}^{\top } =\sigma ^{2}\mathbf{L}(\mathbf{X}_{u}^{\top }\mathbf{X}_{u})^{-1}\mathbf{X}_{u}^{\top }\mathbf{X}_{u}(\mathbf{X}_{u}^{\top }\mathbf{X}_{u})^{-1}\mathbf{L}^{\top } =\sigma ^{2}\mathbf{A}\mathbf{A}^{\top }\). Insbesondere ist also \(\mathbf{A}\mathbf{A}^{\top } = \mathbf{L}(\mathbf{X}_{u}^{\top }\mathbf{X}_{u})^{-1}\mathbf{L}^{\top }\).
- 69.
Es sei y r = X u r aus V r und y a = A ⊤ a aus dem Erzeugnis der Spalten von A ⊤ mit r und a als zugehörigen Koordinatenvektoren bzgl. X u und A ⊤. Mit \(\mathbf{A}^{\top } = \mathbf{X}_{u}(\mathbf{X}_{u}^{\top }\mathbf{X}_{u})^{-1}\mathbf{L}^{\top }\) liegt y a als Linearkombination der Spalten von X u in V u . Da y r in V r liegt, gilt Ly r = 0. Damit folgt \(\mathbf{y}_{a}^{\top }\mathbf{y}_{r} = (\mathbf{A}^{\top }\mathbf{a})^{\top }\mathbf{X}_{u}\mathbf{r} = \mathbf{a}^{\top }\mathbf{A}\mathbf{X}_{u}\mathbf{r} = \mathbf{a}^{\top }\mathbf{L}(\mathbf{X}_{u}^{\top }\mathbf{X}_{u})^{-1}\mathbf{X}_{u}^{\top }\mathbf{X}_{u}\mathbf{r} = \mathbf{a}^{\top }\mathbf{L}\mathbf{r} = \mathbf{a}^{\top }\mathbf{0} = 0\), d. h. y a ⊥ y r . Also liegt y a auch in \(V _{r}^{\perp }\).
- 70.
X 0 = 1, daher ist die Matrix der Residuen E = (I −P 0)Y = QY zentriert (Abschn. 12.1.7, Fußnote 22) und (QY)⊤(QY) deren SSP-Matrix. Als orthogonale Projektion ist Q symmetrisch und idempotent, weshalb \((\mathbf{Q}\mathbf{Y})^{\top }(\mathbf{Q}\mathbf{Y}) = \mathbf{Y}^{\top }\mathbf{Q}^{\top }\mathbf{Q}\mathbf{Y} = \mathbf{Y}^{\top }\mathbf{Q}\mathbf{Y}\) gilt.
- 71.
Zunächst ist die Matrix der Vorhersagedifferenzen \(\hat{\mathbf{Y}}_{u}-\hat{\mathbf{Y}}_{r} = \mathbf{P}_{u}\mathbf{Y}-\mathbf{P}_{r}\mathbf{Y} = \mathbf{Y}-\mathbf{P}_{r}\mathbf{Y}-\mathbf{Y}+\mathbf{P}_{u}\mathbf{Y} = (\mathbf{I}-\mathbf{P}_{r})\mathbf{Y}-(\mathbf{I}-\mathbf{P}_{u})\mathbf{Y} = \mathbf{E}_{r}-\mathbf{E}_{u}\) gleich der Matrix der Differenzen der Residuen. Als Differenz zweier zentrierter Matrizen ist sie damit ihrerseits zentriert. Für die weitere Argumentation s. Fußnote 70.
- 72.
Der gewählte Weg zur Berechnung der Projektionsmatrizen soll die mathematischen Formeln direkt umsetzen, ist aber numerisch nicht stabil und weicht von in R-Funktionen implementierten Rechnungen ab (Bates, 2004).
Literatur
Adler, D., & Murdoch, D. (2016). rgl: 3D visualization device system (OpenGL) [Software]. http://CRAN.R-project.org/package=rgl (R package version 0.96.0)
Adler, J. (2012). R in a Nutshell (2. Aufl.). Sebastopol, CA: O’Reilly.
Agresti, A. (2007). An introduction to categorical data analysis (2. Aufl.). New York, NY: Wiley.
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Thousand Oaks, CA: SAGE.
Allaire, J. J., Cheng, J., Xie, Y., McPherson, J., Chang, W., Allen, J., \(\ldots\) Hyndman, R. (2016). rmarkdown: Dynamic documents for R [Software]. http://rmarkdown.rstudio.com (R package version 1.1)
Allignol, A., & Latouche, A. (2016). CRAN task view: Survival analysis. http://CRAN.R-project.org/view=Survival (Version 2016-08-19)
Allison, P. D. (2008). Convergence failures in logistic regression. SAS Global Forum, 360, 1–11.
Andres, J. (1996). Das allgemeine lineare Modell. In E. Erdfelder, R. Mausfeld, T. Meiser, & G. Rudinger (Hrsg.), Handbuch Quantitative Methoden (S. 185–200). Weinheim: Beltz.
Arnold, J. B. (2016). ggthemes: Extra themes, scales and geoms for ‘ggplot2’ [Software]. https://CRAN.R-project.org/package=ggthemes (R package version 3.2.0)
Backhaus, K., Erichson, B., Plinke, W., & Weiber, R. (2015a). Multivariate Analysemethoden (14. Aufl.). Heidelberg: Springer.
Backhaus, K., Erichson, B., & Weiber, R. (2015b). Fortgeschrittene Multivariate Analysemethoden (3. Aufl.). Heidelberg: Springer.
Balasubramanian, N., & Johnson, S. G. (2013). cubature: Adaptive multivariate integration over hypercubes [Software]. http://CRAN.R-project.org/package=cubature (R package version 1.1-2)
Bates, D. (2004). Least squares calculations in R. R News, 4(1), 17–20. http://CRAN.R-project.org/doc/Rnews/
Bates, D., & Maechler, M. (2016). Matrix: Sparse and Dense Matrix Classes and Methods [Software]. http://CRAN.R-project.org/package=Matrix (R package version 1.2-7.1)
Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1), 1–48. http://www.jstatsoft.org/article/view/v067i01/
Becker, R. A., Chambers, J. M., & Wilks, A. R. (1988). The new S language: A programming environment for data analysis and graphics. Pacific Grove, CA: Wadsworth & Brooks/Cole.
Bender, R., Augustin, T., & Blettner, M. (2005). Generating survival times to simulate Cox proportional hazards models. Statistics in Medicine, 24, 1713–1723.
Bernaards, C. A., & Jennrich, R. I. (2005). Gradient projection algorithms and software for arbitrary rotation criteria in factor analysis. Educational and Psychological Measurement, 65(5), 676–696. http://www.stat.ucla.edu/research/gpa/
Beyersmann, J., Allignol, A., & Schumacher, M. (2012). Competing risks and multistate models with R. New York, NY: Springer.
Bivand, R. S., Pebesma, E., & Gómez-Rubio, V. (2013). Applied spatial data analysis with R. New York, NY: Springer.
Bliese, P. (2016). multilevel: Multilevel Functions [Software]. http://CRAN.R-project.org/package=multilevel (R package version 2.6)
Bloomfield, V. A. (2014). Using R for numerical analysis in science and engineering. Boca Raton, FL: Chapman & Hall/CRC.
Boker, S., Neale, M., Maes, H., Wilde, M., Spiegel, M., Brick, T., \(\ldots\) Fox, J. (2011). OpenMx: An open source extended structural equation modeling framework. Psychometrika, 66(2), 306–317. http://openmx.psyc.virginia.edu/
Borchers, H. W. (2016a). CRAN task view: Numerical mathematics. http://CRAN.R-project.org/view=NumericalMathematics (Version 2016-02-09)
Borchers, H. W. (2016b). pracma: Practical Numerical Math Functions [Software]. http://CRAN.R-project.org/package=pracma (R package version 1.9.5)
Borg, A., & Sariyar, M. (2016). RecordLinkage: Record linkage in R [Software]. https://CRAN.R-project.org/package=RecordLinkage (R package version 0.4-10)
Bortz, J., Lienert, G. A., & Boehnke, K. (2010). Verteilungsfreie Methoden in der Biostatistik (3. Aufl.). Heidelberg: Springer.
Bronstein, I. N., & Semendjajew, K. A. (2012). Springer Taschenbuch der Mathematik (3. Aufl.). Berlin: Springer.
Buerkner, P.-C. (2016). brms: Bayesian regression models using Stan [Software]. https://CRAN.R-project.org/package=brms (R package version 1.0.1)
Büning, H., & Trenkler, G. (1994). Nichtparametrische statistische Methoden (2. Aufl.). Berlin: Walter de Gruyter.
Canty, A., & Ripley, B. D. (2016). boot: Bootstrap R (S-Plus) Functions [Software]. http://CRAN.R-project.org/package=boot (R package version 1.3-18)
Carr, D., Lewin-Koh, N., & Maechler, M. (2015). hexbin: Hexagonal Binning Routines [Software]. http://CRAN.R-project.org/package=hexbin (R package version 1.27.1)
Chambers, J. M. (2008). Software for data analysis: Programming with R. New York, NY: Springer. http://statweb.stanford.edu/~jmc4/Rbook/
Champely, S., & De Rosario, H. (2016). pwr: Basic functions for power analysis [Software]. http://CRAN.R-project.org/package=pwr (R package version 1.2-0)
Chang, W. (2013). R graphics cookbook. Sebastopol, CA: O’Reilly. http://www.cookbook-r.com/
Chang, W., Cheng, J., Allaire, J. J., Xie, Y., & McPherson, J. (2016). shiny: Web application framework for R [Software]. http://CRAN.R-project.org/package=shiny (R package version 0.14.2)
Chang, W., & Luraschi, J. (2016). profvis: Interactive visualizations for profiling R code [Software]. https://rstudio.github.io/profvis/ (R package version 0.3.2)
Chernik, M. R., & LaBudde, R. A. (2011). An introduction to bootstrap methods with applications to R. Hoboken, NJ: Wiley.
Chihara, L., & Hesterberg, T. (2011). Mathematical statistics with resampling and R. Hoboken, NJ: Wiley. https://sites.google.com/site/chiharahesterberg/
Circle Systems. (2015). Stat/Transfer [Software]. http://www.stattransfer.com/ (Version 13)
Cowlishaw, M. F. (2008). Decimal arithmetic FAQ Part 1 – General questions. http://speleotrove.com/decimal/decifaq1.html
Dalgaard, P. (2007). New functions for multivariate analysis. R News, 7(2), 2–7. http://CRAN.R-project.org/doc/Rnews/
Dalgaard, P. (2008). Introductory statistics with R (2. Aufl.). London, UK: Springer. http://publicifsv.sund.ku.dk/~pd/ISwR.html
Davison, A. C., & Hinkley, D. V. (1997). Bootstrap methods and their applications. Cambridge, UK: Cambridge University Press.
De Rosario-Martinez, H. (2015). phia: Post-hoc interaction analysis [Software]. http://CRAN.R-project.org/package=phia (R package version 0.2-1)
Delignette-Muller, M. L., & Dutang, C. (2015). fitdistrplus: An R package for fitting distributions. Journal of Statistical Software, 64(4), 1–34. http://www.jstatsoft.org/v64/i04/
Dowle, M., Short, T., Lianoglou, S., & Srinivasan, A. (2015). data.table: Extension of data.frame [Software]. http://CRAN.R-project.org/package=data.table (R package version 1.9.6)
Duchateau, L., & Janssen, P. (2007). The frailty model. New York, NY: Springer.
Dutang, C. (2016). CRAN task view: Probability distributions. http://CRAN.R-project.org/view=Distributions (Version 2016-08-06)
Eddelbuettel, D. (2013). Seamless R and C++ integration with Rcpp. New York, NY: Springer.
Eddelbuettel, D. (2016). CRAN task view: High-performance and parallel computing with R. http://CRAN.R-project.org/view=HighPerformanceComputing (Version 2016-08-18)
Eid, M., Gollwitzer, M., & Schmitt, M. (2015). Statistik und Forschungsmethoden (5. Aufl.). Weinheim: Beltz.
Faraway, J. J. (2014). Linear models with R (2. Aufl.). Boca Raton, FL: Chapman & Hall/CRC. http://www.maths.bath.ac.uk/~jjf23/LMR/
Faraway, J. J. (2016). Extending the linear model with R: generalized linear, mixed effects and nonparametric regression models (2. Aufl.). Boca Raton, FL: Chapman & Hall/CRC. http://www.maths.bath.ac.uk/~jjf23/ELM/
Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2), 175–191. http://www.gpower.hhu.de/
Filzmoser, P., Fritz, H., & Kalcher, K. (2016). pcaPP: Robust PCA by Projection Pursuit [Software]. http://CRAN.R-project.org/package=pcaPP (R package version 1.9-61)
Filzmoser, P., & Gschwandtner, M. (2015). mvoutlier: Multivariate outlier detection based on robust methods [Software]. http://CRAN.R-project.org/package=mvoutlier (R package version 2.0.6)
Fischer, G. (2013). Lineare algebra (18. Aufl.). Wiesbaden: Vieweg+Teubner.
Fox, J. (2003). Effect displays in R for generalised linear models. Journal of Statistical Software, 8(15), 1–27. http://www.jstatsoft.org/v08/i15/
Fox, J. (2005). The R commander: A basic-statistics graphical user interface to R. Journal of Statistical Software, 14(9), 1–42. http://www.jstatsoft.org/v14/i09/
Fox, J. (2016a). CRAN task view: Statistics for the social sciences. http://CRAN.R-project.org/view=SocialSciences (Version 2016-08-18)
Fox, J. (2016b). polycor: Polychoric and Polyserial Correlations [Software]. http://CRAN.R-project.org/package=polycor (R package version 0.7-9)
Fox, J., Friendly, M., & Weisberg, S. (2013). Hypothesis tests for multivariate linear models using the car package. The R Journal, 5(1), 39–52. https://journal.r-project.org/archive/2013-1/
Fox, J., Nie, Z., & Byrnes, J. (2016). sem: Structural Equation Models [Software]. http://CRAN.R-project.org/package=sem (R package version 3.1-8)
Fox, J., & Weisberg, S. (2011). An R companion to applied regression (2. Aufl.). Thousand Oaks, CA: SAGE. http://socserv.socsci.mcmaster.ca/jfox/Books/Companion/
Fox, J., & Weisberg, S. (2016). car: Companion to Applied Regression [Software]. http://CRAN.R-project.org/package=car (R package version 2.1-3)
Friedman, J., Hastie, T., & Tibshirani, R. (2010). Regularization paths for generalized linear models via coordinate descent. Journal of Statistical Software, 33(1), 1–22. http://www.jstatsoft.org/v33/i01/
Gabry, J., & Goodrich, B. (2016). rstanarm: Bayesian applied regression modeling via Stan [Software]. https://CRAN.R-project.org/package=rstanarm (R package version 2.12.1)
Gałecki, A. T., & Burzykowski, T. (2013). Linear mixed-effects models using R: A step-by-step approach. New York, NY: Springer.
Gandrud, C. (2015). Reproducible research with R & RStudio (2. Aufl.). Boca Raton, FL: Chapman & Hall/CRC. http://christophergandrud.github.io/RepResR-RStudio/
Garnier, S. (2016). viridis: Default color maps from ‘matplotlib’ [Software]. https://CRAN.R-project.org/package=viridis (R package version 0.3.4)
Gentleman, R. C., Carey, V. J., Bates, D. M., Bolstad, B., Dettling, M., Dudoit, S., \(\ldots\) Zhang, J. (2004). Bioconductor: Open software development for computational biology and bioinformatics. Genome Biology, 5, R80. http://genomebiology.com/2004/5/10/R80
Genz, A., & Bretz, F. (2009). Computation of multivariate normal and t probabilities (Lecture notes in statistics, Vol. 195). Heidelberg: Springer.
Genz, A., Bretz, F., Miwa, T., Mi, X., Leisch, F., Scheipl, F., & Hothorn, T. (2016). mvtnorm: Multivariate normal and t distributions [Software]. http://CRAN.R-project.org/package=mvtnorm (R package version 1.0-5)
Gesmann, M., & de Castillo, D. (2011). googleVis: Interface between R and the Google Visualisation API. The R Journal, 3(2), 40–44. http://journal.r-project.org/archive/2011-2/
Gilbert, P., & Varadhan, R. (2016). numDeriv: Accurate numerical derivatives [Software]. https://CRAN.R-project.org/package=numDeriv (R package version 2016.8-1)
Goldberg, D. (1991). What every computer scientist should know about floating-point arithmetic. ACM Computing Surveys, 23(1), 5–48. http://www.validlab.com/goldberg/paper.pdf
Good, P. I. (2004). Permutation, parametric, and bootstrap tests of hypotheses (3. Aufl.). New York, NY: Springer.
Goulet, V., Dutang, C., Maechler, M., Firth, D., Shapira, M., & Stadelmann, M. (2015). expm: Matrix exponential [Software]. http://CRAN.R-project.org/package=expm (R package version 0.999-0)
Goyvaerts, J., & Levithan, S. (2012). Regular expressions cookbook. Sebastopol, CA: O’Reilly. http://www.regular-expressions.info/
Grolemund, G., & Wickham, H. (2011). Dates and Times Made Easy with lubridate. Journal of Statistical Software, 40(3), 1–25. http://www.jstatsoft.org/v40/i03/
Grothendieck, G., & Petzoldt, T. (2004). Date and Time Classes in R. R News, 4(1), 29–32. http://CRAN.R-project.org/doc/Rnews/
Harrell Jr, F. E. (2015). Regression modeling strategies (2. Aufl.). New York: Springer. http://biostat.mc.vanderbilt.edu/wiki/Main/RmS
Harrell Jr, F. E. (2016a). Hmisc: Harrell Miscellaneous [Software]. http://CRAN.R-project.org/package=Hmisc (R package version 4.0-0)
Harrell Jr, F. E. (2016b). rms: Regression Modeling Strategies [Software]. http://CRAN.R-project.org/package=rms (R package version 5.0-0)
Hastie, T., Tibshirani, R., & Friedman, J. (2009). The elements of statistical learning: data mining, inference, and prediction (2. Aufl.). New York: Springer. http://statweb.stanford.edu/~tibs/ElemStatLearn/
Heiberger, R. M., & Neuwirth, E. (2009). R through Excel. New York, NY: Springer.
Heinze, G., Ploner, M., Dunkler, D., & Southworth, H. (2013). logistf: Firth’s bias reduced logistic regression [Software]. http://CRAN.R-project.org/package=logistf (R package version 1.21)
Hendrickx, J. (2012). perturb: Tools for evaluating collinearity [Software]. http://CRAN.R-project.org/package=perturb (R package version 2.05)
Hesterberg, T. (2015). resample: Resampling functions [Software]. http://CRAN.R-project.org/package=resample (R package version 0.4)
Hewson, P. (2015). CRAN task view: Multivariate statistics. http://CRAN.R-project.org/view=Multivariate (Version 2015-07-01)
Hoffman, D. D. (2000). Visual intelligence: How we create what we see. New York, NY: W. W. Norton & Company.
Højsgaard, S., Halekoh, U., & Yan, J. (2006). The R package geepack for generalized estimating equations. Journal of Statistical Software, 15(2), 1–11. http://www.jstatsoft.org/v15/i02/
Honaker, J., King, G., & Blackwell, M. (2011). Amelia: II: A program for missing data. Journal of Statistical Software, 45(7), 1–47. http://www.jstatsoft.org/v45/i07/
Hornik, K. (2016). The R FAQ [Software-Handbuch]. http://CRAN.R-project.org/doc/FAQ/
Hosmer Jr, D. W., Lemeshow, S., & May, S. (2008). Applied survival analysis: Regression modeling of time to event data (2. Aufl.). New York, NY: Wiley.
Hothorn, T. (2016). CRAN task view: Machine learning & statistical learning. http://CRAN.R-project.org/view=MachineLearning (Version 2016-06-24)
Hothorn, T., Bretz, F., & Westfall, P. (2008). Simultaneous inference in general parametric models. Biometrical Journal, 50(3), 346–363.
Hothorn, T., & Everitt, B. S. (2014). A handbook of statistical analysis using R (3. Aufl.). Boca Raton, FL: Chapman & Hall/CRC.
Hothorn, T., & Everitt, B. S. (2015). HSAUR3: A handbook of statistical analyses using R (3rd ed.) [Software]. http://CRAN.R-project.org/package=HSAUR3 (R package version 1.0-5)
Hothorn, T., Hornik, K., van de Wiel, M. A., & Zeileis, A. (2008). Implementing a class of permutation tests: The coin package. Journal of Statistical Software, 28(8), 1–23. http://www.jstatsoft.org/v28/i08/
Huber, W., Carey, V. J., Gentleman, R., Anders, S., Carlson, M., Carvalho, B. S., \(\ldots\) Morgan, M. (2015). Orchestrating high-throughput genomic analysis with Bioconductor. Nature Methods, 12, 115–121.
Hyndman, R. J. (2016). CRAN task view: Time series analysis. http://CRAN.R-project.org/view=TimeSeries (Version 2016-09-09)
Ihaka, R., & Gentleman, R. (1996). R: A language for data analysis and graphics. Journal of Computational and Graphical Statistics, 5(3), 299–314.
Ihaka, R., Murrell, P., Fisher, J. C., & Zeileis, A. (2015). colorspace: Color Space Manipulation [Software]. http://CRAN.R-project.org/package=colorspace (R package version 1.2-6)
James, G., Witten, D., Hastie, T., & Tibshirani, R. (2013). An introduction to statistical learning with applications in R. New York, NY: Springer. http://www-bcf.usc.edu/~gareth/ISL/
Johnson, P. E. (2016). rockchalk: Regression Estimation and Presentation [Software]. http://CRAN.R-project.org/package=rockchalk (R package version 1.8.101)
Johnson, S. G. (2014). The NLopt nonlinear-optimization package [Software]. http://ab-initio.mit.edu/nlopt/ (Version 2.4.2)
Kelley, K. (2016). MBESS [Software]. http://CRAN.R-project.org/package=MBESS (R package version 4.1.0)
Kirk, R. E. (2013). Experimental design – Procedures for the social sciences (4. Aufl.). Thousand Oaks, CA: SAGE. http://www.sagepub.com/kirk/
Klein, J. P., & Moeschberger, M. L. (2003). Survival analysis: Techniques for censored and truncated data (2. Aufl.). New York, NY: Springer.
Kleinman, K., & Horton, N. J. (2014). SAS and R: Data management, statistical analysis, and graphics (2. Aufl.). Boca Raton, FL: Chapman Hall/CRC. http://nhorton.people.amherst.edu/sasr2/
Kolde, R. (2015). pheatmap: Pretty heatmaps [Software]. https://CRAN.R-project.org/package=pheatmap (R package version 1.0.8)
Kruschke, J. K. (2015). Doing Bayesian data analysis: A tutorial with R, JAGS, and Stan (2. Aufl.). Amsterdam: Academic Press.
Kuhn, M. (2016). caret: Classification and regression training [Software]. http://CRAN.R-project.org/package=caret (R package version 6.0-71)
Kuhn, M., & Johnson, K. (2013). Applied predictive modeling. New York, NY: Springer. http://appliedpredictivemodeling.com/
Leeper, T., Chamberlain, S., Mair, P., Ram, K., & Gandrud, C. (2016). CRAN task view: Web technologies and services. http://CRAN.R-project.org/view=WebTechnologies (Version 2016-08-18)
Leisch, F., & Gruen, B. (2016). CRAN task view: Cluster analysis & finite mixture models. http://CRAN.R-project.org/view=Cluster (Version 2016-08-18)
Lemon, J. (2006). Plotrix: a package in the red light district of R. R News, 6(4), 8–12.
Ligges, U. (2002). R Help Desk: Automation of mathematical annotation in plots. R News, 2(3), 32–34. http://CRAN.R-project.org/doc/Rnews/
Ligges, U. (2003). R Help Desk: Package management. R News, 3(3), 37–39. http://CRAN.R-project.org/doc/Rnews/
Ligges, U. (2016). Programmieren mit R (4. Aufl.). Berlin: Springer Spektrum.
Ligges, U., & Fox, J. (2008). R Help Desk: How can I avoid this loop or make it faster? R News, 8(1), 46–50. http://CRAN.R-project.org/doc/Rnews/
Long, J. D. (2012). Longitudinal data analysis for the behavioral sciences using R. Thousand Oaks, CA: SAGE. https://studysites.sagepub.com/long/
Lumley, T. (2009). leaps: Regression subset selection (using Fortran code by Alan Miller) [Software]. http://CRAN.R-project.org/package=leaps (R package version 2.9)
Maechler, M. (2016). CRAN task view: Robust statistical methods. http://CRAN.R-project.org/view=Robust (Version 2016-08-29)
Maindonald, J., & Braun, W. J. (2010). Data analysis and graphics using R: An example-based approach (3. Aufl.). Cambridge, UK: Cambridge University Press. http://maths-people.anu.edu.au/~johnm/r-book/daagur3.html
Maindonald, J., & Braun, W. J. (2015). DAAG: Data Analysis And Graphics data and functions [Software]. http://CRAN.R-project.org/package=DAAG (R package version 1.22)
Mair, P. (2016). CRAN task view: Psychometric models and methods. http://CRAN.R-project.org/view=Psychometrics (Version 2016-08-18)
Mardia, K. V., Kent, J. T., & Bibby, J. M. (1980). Multivariate Analysis. London, UK: Academic Press.
Maronna, R. A., Martin, R. D., & Yohai, V. J. (2006). Robust Statistics: Theory and Methods. New York, NY: Wiley.
Maxwell, S. E., & Delaney, H. D. (2004). Designing experiments and analyzing data: A model comparison perspective (2. Aufl.). Mahwah, NJ: Lawrence Erlbaum.
Mazerolle, M. J. (2016). AICcmodavg: Model selection and multimodel inference based on (Q)AIC(c) [Software]. http://CRAN.R-project.org/package=AICcmodavg (R package version 2.0-4)
McElreath, R. (2015). Statistical rethinking: A Bayesian course with examples in R and Stan. Boca Raton, FL: Chapman & Hall/CRC.
McFarlane, J. (2016). pandoc: A universal document converter [Software]. http://pandoc.org/ (Version 1.18)
Mersmann, O. (2015). microbenchmark: Sub microsecond accurate timing functions [Software]. http://CRAN.R-project.org/package=microbenchmark (R package version 1.4-2.1)
Meyer, D., & Hornik, K. (2009). Generalized and customizable sets in R. Journal of Statistical Software, 31(2), 1–27. http://www.jstatsoft.org/v31/i02/
Meyer, D., Zeileis, A., & Hornik, K. (2016). vcd: Visualizing Categorical Data [Software]. http://CRAN.R-project.org/package=vcd (R package version 1.4-3)
Meyer, S., & Held, L. (2014). Power-law models for infectious disease spread. The Annals of Applied Statistics, 8(3), 1612–1639.
Microsoft. (2016). checkpoint: Install packages from snapshots on the checkpoint server for producibility [Software]. https://CRAN.R-project.org/package=checkpoint (R package version 0.3.18)
Microsoft. (2016a). Microsoft R Open [Software]. http://mran.revolutionanalytics.com/open/ (3.3.1)
Microsoft. (2016b). R Tools for Visual Studio [Software]. https://www.visualstudio.com/vs/rtvs/ (Version 0.5)
Miller, A. J. (2002). Subset selection in regression (2. Aufl.). Boca Raton, FL: Chapman & Hall/CRC.
Muenchen, R. A. (2011). R for SAS and SPSS users (2. Aufl.). New York, NY: Springer. http://r4stats.com/
Muenchen, R. A., & Hilbe, J. M. (2010). R for Stata users. New York, NY: Springer. http://r4stats.com/
Murrell, P. (2011). R Graphics (2. Aufl.). Boca Raton, FL: Chapman & Hall/CRC. http://www.stat.auckland.ac.nz/~paul/RG2e/
Nash, J. C. (2014a). Nonlinear parameter optimization using R tools. Chichester, UK: Wiley.
Nash, J. C. (2014b). On best practice optimization methods in R. Journal of Statistical Software, 60(2), 1–14. http://www.jstatsoft.org/v60/i02/
Nash, J. C., & Ravi, V. (2011). Unifying optimization algorithms to aid software system users: Optimx for R. Journal of Statistical Software, 43(9), 1–14. http://www.jstatsoft.org/v43/i09/
Neuwirth, E. (2014). RColorBrewer: ColorBrewer palettes [Software]. http://CRAN.R-project.org/package=RColorBrewer (Version 1.1-2)
Neuwirth, E. (2016). RExcel [Software]. http://rcom.univie.ac.at/ (Version 3.2.16)
Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, P., Minchin, P. R., O’Hara, R. B., \(\ldots\) Wagner, H. (2016). vegan: Community Ecology Package [Software]. http://CRAN.R-project.org/package=vegan (R package version 2.4-1)
OpenAnalytics BVBA. (2015). Architect [Software]. http://www.openanalytics.eu/architect/ (Version 0.9.8)
Park, J. H. (2016). CRAN task view: Bayesian inference. http://CRAN.R-project.org/view=Bayesian (Version 2016-09-19)
Pau, G., Fuchs, F., Sklyar, O., Boutros, M., & Huber, W. (2010). EBImage – an R package for image processing with applications to cellular phenotypes. Bioinformatics, 26(7), 979–981. http://www.bioconductor.org/packages/devel/bioc/html/EBImage.html
Peters, A., & Hothorn, T. (2015). ipred: Improved predictors [Software]. http://CRAN.R-project.org/package=ipred (R package version 0.9-5)
Pinheiro, J. C., & Bates, D. M. (2000). Mixed-effects models in S and S-PLUS. New York, NY: Springer.
Pinheiro, J. C., Bates, D. M., DebRoy, S., Sarkar, D., & R Core Team. (2016). nlme: Linear and Nonlinear Mixed Effects Models [Software]. http://CRAN.R-project.org/package=nlme (R package version 3.1-128)
Plate, T., & Heiberger, R. (2016). abind: Combine multi-dimensional arrays [Software]. http://CRAN.R-project.org/package=abind (R package version 1.4-5)
Ploner, M., & Heinze, G. (2015). coxphf: Cox regression with Firth’s penalized likelihood [Software]. http://CRAN.R-project.org/package=coxphf (R package version 1.11)
Polzehl, J., & Tabelow, K. (2007). Adaptive Smoothing of Digital images: The R Package adimpro. Journal of Statistical Software, 19(1), 1–17. http://www.jstatsoft.org/v19/i01/
Poncet, P. (2012). modeest: Mode estimation [Software]. http://CRAN.R-project.org/package=modeest (R package version 2.1)
R Core Team. (2014). R: A language and environment for statistical computing [Software-Handbuch]. Vienna, Austria. http://www.r-project.org/
R Core Team. (2016a). foreign: Read data stored by Minitab, S, SAS, SPSS, Stata, Systat, dBase, …[Software]. http://CRAN.R-project.org/package=foreign (R package version 0.8-67)
R Core Team. (2016b). R: Data Import/Export [Software-Handbuch]. Vienna, Austria. http://CRAN.R-project.org/doc/manuals/R-data.html
R Core Team. (2016c). R: Installation and Administration [Software-Handbuch]. Vienna, Austria. http://CRAN.R-project.org/doc/manuals/R-admin.html
R Core Team. (2016d). Writing R Extensions [Software-Handbuch]. Vienna. http://cran.r-project.org/doc/manuals/R-exts.html
R Foundation for Statistical Computing. (2014). R: Regulatory compliance and validation issues – A guidance document for the use of R in regulated clinical trial environments [Software-Handbuch]. Vienna, Austria. http://www.r-project.org/doc/R-FDA.pdf
R Special Interest Group on Databases, Wickham, H., & Müller, K. (2016). DBI: R database interface [Software]. http://CRAN.R-project.org/package=DBI (R package version 0.5-1)
Revelle, W. (2016). psych: Procedures for Psychological, Psychometric, and Personality Research [Software]. http://CRAN.R-project.org/package=psych (R package version 1.6.9)
Ripley, B. D. (2016). RODBC: ODBC Database Access [Software]. http://CRAN.R-project.org/package=RODBC (R package version 1.3-14)
Ripley, B. D., & Murdoch, D. (2016). R for Windows FAQ. http://CRAN.R-project.org/bin/windows/base/rw-FAQ.html
Ritz, C., & Streibig, J. C. (2009). Nonlinear regression with R. New York, NY: Springer.
Robin, X., Turck, N., Hainard, A., Tiberti, N., Lisacek, F., Sanchez, J.-C., & Müller, M. (2011). pROC: an open-source package for R and S+ to analyze and compare ROC curves. BMC Bioinformatics, 12, 77. http://web.expasy.org/pROC/
Rödiger, S., Friedrichsmeier, T., Prasenjit, K., & Michalke, M. (2011). RKWard: A Comprehensive Graphical User Interface and Integrated Development Environment for Statistical Analysis with R. Journal of Statistical Software, 49(9), 1–34. http://www.jstatsoft.org/v49/i09/
Rosseel, Y. (2012). lavaan: An R Package for Structural Equation Modeling. Journal of Statistical Software, 48(2), 1–36. http://www.jstatsoft.org/v48/i02/
Rossini, A. J., Maechler, M., Sparapani, R. A., Eglen, S., Spinu, V., & Heiberger, R. M. (2016). ESS: Emacs Speaks Statistics [Software]. http://ess.r-project.org/ (Version 16.04)
Rousseeuw, P. J., Croux, C., Todorov, V., Ruckstuhl, A., Salibian-Barrera, M., Verbeke, T., & Maechler, M. (2016). robustbase: Basic Robust Statistics [Software]. http://CRAN.R-project.org/package=robustbase (R package version 0.92-6)
Rousseeuw, P. J., & van Zomeren, B. C. (1990). Unmasking Multivariate Outliers and Leverage Points. Journal of the American Statistical Association, 85(411), 633–639. http://www.jstor.org/stable/2289995
RStudio Inc. (2016a). ggvis: Interactive grammar of graphics [Software]. http://CRAN.R-project.org/package=ggvis (R package version 0.4.3)
RStudio Inc. (2016b). RStudio [Software]. http://www.rstudio.org/ (Version 1.0.39)
Shumway, R. H., & Stoffer, D. S. (2016). Time Series Analysis and Its Applications (4. Aufl.). New York, NY: Springer. http://www.stat.pitt.edu/stoffer/tsa4/
Sievert, C., Parmer, C., Hocking, T., Chamberlain, S., Ram, K., Corvellec, M., & Despouy, P. (2016). plotly: Create interactive web graphics via Plotly’s JavaScript graphing library [Software]. https://CRAN.R-project.org/package=plotly (R package version 4.5.2)
Signorell, A. (2016). DescTools: Tools for descriptive statistics. http://CRAN.R-project.org/package=DescTools (R package version 0.99.17)
Smyth, G., Hu, Y., Dunn, P., Phipson, B., & Chen, Y. (2016). statmod: Statistical modeling [Software]. https://CRAN.R-project.org/package=statmod (R package version 1.4.26)
Soetaert, K. (2015). rootSolve: Nonlinear root finding, equilibrium and steady-state analysis of ordinary differential equations [Software]. https://CRAN.R-project.org/package=rootSolve (R package version 1.6.6)
Soetaert, K., & Petzoldt, T. (2015). CRAN task view: Differential equations. http://CRAN.R-project.org/view=DifferentialEquations (Version 2015-07-03)
Spector, P. (2008). Data manipulation with R. New York, NY: Springer.
Strang, G. (2003). Lineare Algebra. Berlin: Springer.
Su, Y.-S., Gelman, A., Hill, J., & Yajima, M. (2011). Multiple imputation with diagnostics (mi) in R: Opening windows into the black box. Journal of Statistical Software, 45(2), 1–67. http://www.jstatsoft.org/v45/i02/
Templ, M. (2016). CRAN task view: Official statistics & survey methodology. http://CRAN.R-project.org/view=OfficialStatistics (Version 2016-06-28)
Therneau, T. (2016). survival: Survival analysis [Software]. http://CRAN.R-project.org/package=survival (R package version 2.39-5)
Theussl, S., & Borchers, H. W. (2016). CRAN task view: Optimization and mathematical programming. http://CRAN.R-project.org/view=Optimization (Version 2016-06-06)
TIBCO Software Inc. (2016). TIBCO enterprise runtime for R (TERR) [Software]. http://spotfire.tibco.com/ (Version 4.1)
Tingley, D., Yamamoto, T., Keele, L., & Imai, K. (2015). mediation: R Package for Causal Mediation Analysis [Software]. http://CRAN.R-project.org/package=mediation (R package version 4.4.5)
Unwin, A. (2015). Graphical data analysis with R. Boca Raton, FL: Chapman & Hall/CRC. http://www.gradaanwr.net/
Urbanek, S., & Horner, J. (2015). Cairo: R graphics device using cairo graphics library for creating high-quality bitmap (PNG, JPEG, TIFF), vector (PDF, SVG, PostScript) and display (X11 and Win32) output [Software]. http://CRAN.R-project.org/package=Cairo (R package version 1.5-9)
Ushey, K., McPherson, J., Cheng, J., Atkins, A., & Allaire, J. (2016). packrat: A dependency management system for projects and their R package pendencies [Software]. https://CRAN.R-project.org/package=packrat (R package version 0.4.8-1)
van Buuren, S. (2012). Flexible imputation of missing data. Boca Raton, FL: Chapman & Hall/CRC.
van Buuren, S., & Groothuis-Oudshoorn, K. (2011). MICE: Multivariate Imputation by Chained Equations in R. Journal of Statistical Software, 45(3), 1–67. http://www.jstatsoft.org/v45/i03/
van der Loo, M. P. J. (2014). The stringdist package for approximate string matching. The R Journal, 6,111–122. http://CRAN.R-project.org/package=stringdist
Venables, W. N. (2016). codingMatrices: Alternative factor coding matrices for linear model formulae [Software]. https://CRAN.R-project.org/package=codingMatrices (R package version 0.2.1)
Venables, W. N., & Ripley, B. D. (2002). Modern applied statistics with S (4. Aufl.). New York, NY: Springer. http://www.stats.ox.ac.uk/pub/MASS4/
Venables, W. N., Smith, D. M., Gentleman, R., Ihaka, R., Maechler, M., & R Core Team. (2016). An introduction to R [Software-Handbuch]. Vienna. http://CRAN.R-project.org/doc/manuals/R-intro.html
Wahlbrink, S. (2016). Eclipse Plug-In for R: StatET [Software]. http://www.walware.de/goto/statet (Version 3.5.1)
Walker, A. (2015). openxlsx: Read, write and edit XLSX files [Software]. http://CRAN.R-project.org/package=openxlsx (R package version 3.0.0)
Wand, M. (2015). KernSmooth: Functions for kernel smoothing for Wand & Jones (1995) [Software]. http://CRAN.R-project.org/package=KernSmooth (R package version 2.23-15j)
West, B. T., Welch, K. B., & Gałecki, A. T. (2014). Linear mixed models: A Practical guide using statistical software (2. Aufl.). Boca Raton, FL: Chapman & Hall/CRC. http://www-personal.umich.edu/~agalecki/
Wickham, H. (2014). Advanced R programming. Boca Raton, FL: Chapman & Hall/CRC. http://adv-r.had.co.nz/
Wickham, H. (2015). R packages. Sebastopol, CA: O’Reilly. http://r-pkgs.had.co.nz/
Wickham, H. (2016a). devtools: Tools to make developing R code easier [Software]. http://CRAN.R-project.org/package=devtools (R package version 1.12.0)
Wickham, H. (2016b). readxl: Read Excel files [Software]. http://CRAN.R-project.org/package=readxl (R package version 0.1.1)
Wickham, H. (2016c). stringr: Simple, consistent wrappers for common string operations [Software]. http://CRAN.R-project.org/package=stringr (R package version 1.1.0)
Wickham, H. (2016d). tidyr: Easily tidy data with spread and gather functions [Software]. http://CRAN.R-project.org/package=tidyr (R package version 0.6.0)
Wickham, H. (2016e). tidyverse: Easily install and load ‘tidyverse’ packages [Software]. https://CRAN.R-project.org/package=tidyverse (R package version 1.0.0)
Wickham, H., Cook, D., Hofmann, H., & Buja, A. (2011). tourr: An R package for exploring multivariate data with projections. Journal of Statistical Software, 40(2), 1–18. http://www.jstatsoft.org/v40/i02/
Wickham, H., & Francois, R. (2016a). dplyr: A grammar of data manipulation [Software]. http://CRAN.R-project.org/package=dplyr (R package version 0.5.0)
Wickham, H., & Francois, R. (2016b). readr: Read tabular data [Software]. http://CRAN.R-project.org/package=readr (R package version 1.0.0)
Wickham, H., & Grolemund, G. (2016). R for data science. Sebastopol, CA: O’Reilly. http://r4ds.had.co.nz/
Wickham, H., James, D. A., Falcon, S., SQLite Authors, Healy, L., & RStudio Inc. (2014). RSQLite: SQLite interface for R [Software]. http://CRAN.R-project.org/package=RSQLite (R package version 1.0.0)
Wickham, H., & Miller, E. (2016). haven: Import SPSS, Stata and SAS files [Software]. http://CRAN.R-project.org/package=haven (R package version 1.0.0)
Wickham, H., & Sievert, C. (2016). ggplot2: Elegant graphics for data analysis (2. Aufl.). New York, NY: Springer. http://github.com/hadley/ggplot2-book
Wilke, C. O. (2016). cowplot: Streamlined plot theme and plot annotations for ggplot2 [Software]. https://CRAN.R-project.org/package=cowplot (R package version 0.6.3)
Wilkinson, G. N., & Rogers, C. E. (1973). Symbolic description of factorial models for analysis of variance. Applied Statistics, 22, 392–399.
Wilkinson, L. (2005). The Grammar of Graphics (2. Aufl.). New York, NY: Springer.
Wirtz, M., & Caspar, F. (2002). Beurteilerübereinstimmung und Beurteilerreliabilität. Göttingen: Hogrefe.
Wood, S. N. (2006). Generalized additive models: An introduction with R. Boca Raton, FL: Chapman & Hall/CRC.
Würtz, D., Chalabi, Y., & Maechler, M. (2015). timeDate: Rmetrics – Chronological and Calendar Objects [Software]. http://CRAN.R-project.org/package=timeDate (R package version 3012.100)
Xie, Y. (2015). Dynamic documents with R and knitr (2. Aufl.). Boca Raton, FL: Chapman & Hall/CRC.
Xie, Y. (2016). knitr: A general-purpose package for dynamic report generation in R [Software]. http://CRAN.R-project.org/package=knitr (R package version 1.14)
Yan, J., & Fine, J. P. (2004). Estimating equations for association structures. Statistics in Medicine, 23(6), 859–874.
Yee, T. W. (2010). The VGAM package for categorical data analysis. Journal of Statistical Software, 32(10), 1–34. http://www.jstatsoft.org/v32/i10/
Zeileis, A. (2004). Econometric computing with HC and HAC covariance matrix estimators. Journal of Statistical Software, 11(10), 1–17. http://www.jstatsoft.org/v11/i10/
Zeileis, A. (2005). CRAN task views. R News, 5(1), 39–40. http://CRAN.R-project.org/doc/Rnews/
Zeileis, A., & Hothorn, T. (2002). Diagnostic checking in regression relationships. R News, 2(3), 7–10. http://CRAN.R-project.org/doc/Rnews/
Zeileis, A., Kleiber, C., & Jackman, S. (2008). Regression models for count data in R. Journal of Statistical Software, 27(8), 1–25. http://www.jstatsoft.org/v27/i08/
Zelterman, D. (2015). Applied multivariate statistics with R. Cham: Springer.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2017 Springer-Verlag GmbH Deutschland
About this chapter
Cite this chapter
Wollschläger, D. (2017). Multivariate Verfahren. In: Grundlagen der Datenanalyse mit R. Statistik und ihre Anwendungen. Springer Spektrum, Berlin, Heidelberg. https://doi.org/10.1007/978-3-662-53670-4_12
Download citation
DOI: https://doi.org/10.1007/978-3-662-53670-4_12
Published:
Publisher Name: Springer Spektrum, Berlin, Heidelberg
Print ISBN: 978-3-662-53669-8
Online ISBN: 978-3-662-53670-4
eBook Packages: Life Science and Basic Disciplines (German Language)