This is a collection of R scripts for my personal, research purposes. I make no claims about support or even documentation. If anyone chooses to use this, please do so at your own peril. I am not responsible for any unseemly outcome.
Brief Descriptions:
kmeanssampler.R: Simulate a 2D Gaussian Mixture Model, estimate k-means from it and plot results.
testargs.R: Script to read command line arguments into R scripts
svm_viz.R: Generating a toy GMM and applying SVM and visualizing it.
simple-sir.R: Simulates and inspects a simple 3 variable compartmental model (SIR model) to model for social contagion
pca-iris.R: Uses Fisher's Iris Dataset to perform Principal Component Analysis (PCA) and its diagnoistics.
LMM_Examples.R: A simple implementation of a linear mixed model. These types of models are also known as Hierarchical Linear Models or Random Effects Models.
kmeans.R: A script which takes example data from kmeans-data and then runs the kmeans algorithm on it. Easy and fast - can be modified for any simple dataset's purpose.
structuralequation-examplemodels.R: A script which fits a standard structural equation model on sample data (provided within the script)
t-test-robustness.R: A script which simulates some data and inspects the robustness of the standard 2 sample t-test.
small-multiple-example.R: Applies Tufte's small multiple principle to data analysis with many categories to make visualization and analysis simpler and more palatable. The example output is in small-multiple-example.png coin-fairness-bayesiantest.R: A script which simulates a series of coin tosses and tests the hypothesis that the coin is a fair coin.
robust-regressions.R: A script which performs robust multiple linear regression on data. This type of a regression is usually quite flexible from small departures from the usual assumptions.
bayesian-sensitivityanalysis.R: A script which takes some sample data and performs Bayesian sensitivity analysis on the results of an experiment. This is slightly different from the usual frequentist way of doing sensitivity analysis.
proportion-comparing.R: A script which makes up some fake data and statistically tests between 2 proportions from a bayesian perspective. This type of analysis is very quick, dirty and useful in hci research.
hierarchicalmodeling-bayesian.R: A script which performs randomized hierarchical, bayesian modeling.
normalmean-knownvarian-bayesian.R: A script which simulates a gaussian mean with known variances and looks at its moments.