Bayesian statistics has been around for a few hundred years. Bayesian networks is a much newer branch of mathematics that combines Bayesian statistics with graph theory. Pioneered in the 1980s by 2011 Turing Award Winner Judea Pearl, Bayesian networks can be used to infer causality directly from observational data without the need for a randomized [..]

Data refers to the actual contents of the dataset or the individual data points (e.g., total sales for a product X on date Y).

A specific instance of data or a file within a particular source and format. For example, there might be multiple datasets (one for each year of data) in the same data format for point of sale data.

The data format is the specific structure of the data, meaning the order of the columns and rows.

A particular category or type of data, like social media data or macroeconomic data. There can multiple datasets or formats within one source, since there may be more than one method of tracking data for that category.

We see a wide variety of data to analyze. These datasets could be customer purchases, oil prices or error messages from an engine. Sometimes we can analyze the data as is. Many times however, the raw data may not be easy to analyze due to variations within the data or because the units of one variable may not be easily comparable to units of anothe [..]

This is a visual presentation of a Bayesian (or casual) network model. It will show a visual map of variables, how they relate to each other and the outcome variable.

Once you build a predictive model, it’s a lot more interesting to ask what it would predict in a new situation that’s never been seen before. It turns out that all algorithms, by their nature, perform better on data that they have seen before (known as in-sample or training data), versus data they have never seen before (known as out-of-sample data [..]

Most statistical models are parametric. That is, there is a simple mathematical relationship that describes how to generate new data and those rules involve some numbers. These numbers are known as parameters. The task of model building is to choose rules and parameters that work well on data that we have never seen before. In Bayesian statistics, [..]

Specifies the confidence one has in a given hypothesis. A typical example would be a polling result that is “correct to 2 percentage points, 19 times out of 20.” This is a 95% significance level. What it states is that, under the assumptions underlying the model, if we repeated the poll a large number of times (say 1000), that we expect that the di [..]