NEWNOTES.txt

Absorbance= c(0.084, 0.183, 0.326, 0.464, 0.643, 0.707, 0.717, 0.734 ,0.749 ,0.732) ;
Concentration= c(0.123, 0.288, 0.562, 0.921, 1.420, 1.717, 1.921, 2.137 ,2.321, 2.467) ;
plot(Concentration,Absorbance,pch=18,col="red",font.axis=2,font.lab=2)
abline(coef(lm(Absorbance~Concentration)))

Conc.Squared = (Concentration^2)
Conc.Cubed = (Concentration^3)
ModelA = lm(Absorbance~Concentration)
ModelB = lm(Absorbance~Concentration+Conc.Squared)
ModelC = lm(Absorbance~Concentration+Conc.Squared+Conc.Cubed)



%=============================================================================%


Regression 
Count Modelling
Statistical Disclosure Control
The boot package bootstrapping techniques with R

Rcasts - Stats Inference - Stats Infer 2 - updated


Prague Plans Free Tour 10-14 Castle Tour 14:30
Python Survival Analysis
Pythin Logistic Regression yhat

%=======================================================%
Let $G$ be a simple graph. Explain why the sum of the degrees of the vertices of $G$ is twice the number of its edges.

%=======================================================%

r environment, environment r, introduction to r lab, r programming, environment in r



http://www.scala-lang.org/download/

%==========================================%

%- http://stackoverflow.com/questions/25438899/using-values-in-a-column-to-create-a-ggvis-checkbox

% Image 1

%==========================================%

Numerical and Scientific Computing with Python

\begin{itemize}
\item Matrices
\end{itemize}

%============================================================%

Pareto
Twos Complement
General Term of Maclaurin Series
Ratio Test



MS4024 - PART E Exercise - Assign Data Sets
ggVis Workshop - April 22nd  - Check London R for details
MA4702 - End of Semester Exam Paper - Formula
Limerick cine club
WIDS - Gaby Valdez 
MA4702 Tutorials
MCS
MS4024
dplyr
geraldine idiro
%============================%
% 2006 
$y = x^2 – 6x + 5$

$y = x – 5$

Let both equations equal each other

\[ x^2 – 6x + 5 = x-5 \]

Simplify the expression and solve it.
\[ x^2 - 7x + 10 = 0 \]

\[ (x - 2)(x -5) = 0 \]

Therefore $x=2$ and $x=5$. So the lower and upper bounds of the definite integral are $a=2$ and $b=5$ respectively.

Necessarily these are the lower and upper bounds of the integral.

\[ \int^5_2  - x^2 + 7x - 10 = 0 \]

%===========================================================%
Show the code that would
\begin{itemize}

\item Provide a mean and standard deviation of the Petal Widths for each of the three species of iris.

\item Produce a subset of the data set that only includes the setosa and virginica species.

\item Sort the iris data set in ascending order of Petal Length, then by descending order of Sepal Length.

\item Create an Additional Column that would contain the ratio of Petal Widths to Sepal Widths 
\end{itemize}
%==========================================================%

%============================%
\begin{frame}
\frametitle{The Data Science Profession}
\begin{quote}
MOOC have not  decreased the barrier of entry to machine-learning.
 
Nowadays, you cannot be 'the guy who knows how to run (insert off-the-shelf-algo-here)'. 

In dataland, that's the equivalent to being a code monkey. MOOCs and superb libraries (scikit-learn, R's ecosystem) made 
sure there is plenty of people who can throw say a random forest to a problem. In the modern world, this is not adding that much value. 
\end{quote}
\end{frame}
%====================================================%

Maclaurin Series

(a) Find the Maclaurin series of f(x) = cos(x)  up to and including the term 
containing 
$x^4$
 Use your answer to estimate the value of cos(1)

(a) Find the Maclaurin series of $f(f) = e^x$ 
 up to and including the term containing 



 Use your answer to estimate the value of $e^{0.2}$ and $e^3$


%===========================================%


For Loop

Even Numbers 2n
Odd Numbers 2n-1
cos(n \pi) = -1^{(n)}
sin(n \pi) = -1^{(n+1)}
%========================================================================%
int
diff
Dsolve
quad
polyder
poly and roots
ode23
ode45
rref


%=================================================================================%


% - http://www.nss.gov.au/nss/home.nsf/pages/Confidentiality+-+Managing+the+risk+of+disclosure+in+the+release+of+microdata
% - Types of disclosure risk in microdata


\begin{itemize}
\item R command and R object - Typewriter Font
\item R Package name - Italics
\item Selected Accroymns - Italics
\end{itemize}

%===================================================================%
\textit{
The previous Section (Section 4) is a literary review of residual diagnostics and influence procedures
for Linear Mixed Effects Models, drawing heavily on Schabenberger and Zewotir.

Section 4 begins with an introduction to key topics in residual diagnostics, such as influence, leverage, outliers
and Cook's distance. Other concepts such as \textit{DFFITS} and \textit{DFBETA}s will be introduced briefly, mostly to explain why the are not particularly useful for
the Method Comparison context, and therefore are not elaborated upon.

In brief, Variable Selection is not applicable to Method Comparison Studies, in the 
commonly used used context. 
Testing a rather simplisticy specificied model against one with more random effects terms is tractable, but this research question is of secondary importance.

}

\newpage
%====================================================================%

\subsection{Appendix to Section 4}

As an appendix to section 4, an appraisal of the current state of development (or lack thereof) for current implemenations for LME models, particularly for
\texttt{nlme} and \texttt{lme4} fitted models.

Crucially, a review of internet resources indicates that almost all of the progress in this regard has been done for \texttt{lme4} fitted models, specifically the \textit{Influence.ME} \texttt{R} package. CITATION.
Conversely there is very little for \texttt{nlme} models. One would immediately look at the current development workflow for both packages.
 
%======================%
% Douglas Bates

As an aside, Douglas Bates was arguably the most prominent \texttt{R} developer working in the LME area. 
However Bates has now prioritised the development of LME models in another computing environment , i.e Julia. 
% The current version of this is XXXX

%======================%
% nlme

With regards to \texttt{nlme}, the torch has been passed to Galecki et al (UMich.).  Galecki & Burzykowski published Linear Mixed Effects Models using \texttt{R}. 
Also, the accompanying \texttt{R} package, nlmeU package is under current development, with a version being released XXXX.


%======================%
% lme4 and influence.ME


The \texttt{lme4} package is also under active development, under the leadership of Ben Bolker (UChig.).


%=====================%
% Important Consideration for MCS

The key issue is that \texttt{nlme} allows for the particular specification of Roy's Model, speciifically direct spefiication of the VC matrices for within subject and between subject residuals.
The \texttt{lme4} package does not allow for this.
To advance the ideas that eminate from Roys' paper, one is required to use the \texttt{nlme} context. However, to take advantage of the infrastructure already provided for \texttt{lme4} models, one may change the research question away from that of Roy's paper. 
To this end, an exploration of what textit{influence.ME} can accomplished is merited.

%====================%
% Diagnostics with nlmeU

\subsection*{Leave-One-Out Diagnostics with \texttt{lmeU}}
Galecki et al discuss the matter of LME influence diagnostics in their book.

The command \texttt{lmeU} fits a model with a particular subject removed. The identifier of the subject to be removed is passed as the only argument

A plot ofthe per-observation diagnostics individual subject log-likelihood contributions can be rendered.

\subsubsection*{Likelihood Displacement}
%% Page 503 Galecki


\newpage
%====================================================================%



%============================================================================%

Measurement System Analysis 
Book Six Sigma with R : Statistical Engineering for Process Improvement

Cano, Moguerza, Redchuk


The Table 5.1 of SixSigma book is noticeable in the data can paralleled with carstensens' formulation

\begin{itemize}
\item Voltmeter : Method 
\item Item : battery
\item Replicate : Run
\item y : voltage
\end{itemize}


\begin{framed}
\begin{Verbatim}

\end{verbatim}
\end{Framed}
%===========================================================================%
%===========================================================================%

\subsection{Linear Mixed Effects Models}

Linear Model : Fixed Effects Predictors Only

Linear Mixed Effects Models : Fixed Effects and Random Effects

%-----------------------------------------------------%
\subsection{Fixed Effects and Random Effects}


%-----------------------------------------------------%
\subsection*{Implementations with \texttt{R}} 

\begin{description}
\item[nlme]

\item[lme4]

\end{description}
%=====================================================%

Can you use Bootstrapping for LMEs?

 - http://support.sas.com/resources/papers/proceedings14/1558-2014.pdf


https://projecteuclid.org/download/pdf_1/euclid.ss/1032280214



R in Insurance

%===================================================%
\begin{frame}
\frametitle{Predicting a count variable }
\[ \hat{y}_1 = e^{} \times e^{ \times 1 } \times e^{ \times 0} \times e^{ \times 60} \]


\[ \hat{y}_2 = e^{} \times e^{ \times 0} \times e^{ \times 1 } \times e^{ \times 60} \]


\[ \hat{y}_3 = e^{} \times e^{ \times 0 } \times e^{ \times 0 } \times e^{ \times 60} \]
\end{frame}
%=====================================================%

http://www.unmotivating.com/16-of-the-most-magnificent-trees-in-the-world/

The universe sent us a smiley face from 7.5 billion light-years away :) 
http://slate.me/1DVX9dM  

Neglect of Probability
p-values
Monty Hall Problem
Gambler's Ruin
Canadian Car Lottery Winner
Perugia Murders Amanda Knox
Public Engagement


%===================================================%

hamilton
coding grace /pydata **
meetup.com
FB
seaborn
pat dicker**
Mick Cooney**
intercom
UL**
HibColl
install ipython


lmer
intercom

Mandatory dependencies
numpy
scipy
matplotlib
pandas
Recommended dependencies
statsmodels
patsy