Services Offered

The Applied Information Company provides state-of-the-science courses in the “information-theoretic” (I-T) paradigm for data analysis and inference. These 2- or 3-day courses focus on a new paradigm for statistical data analysis. Read More....

Costs and Related Details

The costs depend on several factors but an example will provide a rough guideline as to what to expect. Consider a 2-day course within the USA where 30+ go to register... Read More....

Why This Course Is Superior

Many technical people are tired of statistical analysis methods that involve a test of a (usually trivial) null hypothesis, a test statistic, an α-level, a P-value, and an arbitrary ruling concerning its “statistical significance.” Read More....

Information Theory Seminars & Workshops

Background Required For Attendees

Ideally, attendees of our information theory seminars and workshops should have some experience with Fisher’s likelihood (e.g., logistic regression) concepts such as the likelihood function, log-likelihood function, maximum likelihood estimation, deviance, profile likelihood intervals) would be nice. (I understand that people may have little knowledge of likelihood issues).

With the above background, people will find the quantitative material easy; it is the philosophical and conceptual issues that are challenging to nearly everyone.

These information theory workshops are highly interactive and learning is fun. The days will go by quickly. The information theory course is ideally suited for a person that has thought hard about the context of the issue, developed several sound alternative hypotheses, derived a model for each hypothesis, collected data, and obtained the MLEs for the model parameters in each model and their covariance matrices.

Then the focus is on “which model is best to use for inferences” (e.g., prediction).

No need to bring a laptop except participants will need a calculator or computer for the hand on session the second morning of the course (+, - , ×, ÷ , loge.exp).

Disclaimer This is not a modeling course. I will showcase a variety of models and attendees will gain some insights into modeling, but the short course does not focus on model building or the estimation of model parameters (via maximum likelihood or least squares). In addition, little is said about various random effects models and I say little about Bayesian approaches.

Models well outside the likelihood framework are not covered (e.g., CART methods). I will make a few comments about model selection using generalized estimating equations. Lastly I do not deal with various multivariate methods as these seem to be rare in general usage.