1294
Part IX: Business Intelligence
Logistic Regression
Logistic regression is a special case of the neural network algorithm. Although you can use
logistic regression for many tasks, it is especially suited for estimation problems where lin-
ear regression would be a good fi t. However, because the predicted value is discrete,
the linear approach tends to predict values outside the allowed range — for example, pre-
dicting probabilities more than 100 percent for a certain combination of inputs.Because it is derived from the neural network algorithm, logistic regression shares the same
viewer.Naive Bayes
Naive Bayes is a fast algorithm with accuracy adequate for many applications. It does not,
however, operate on continuous variables. Every input is independent. For example, the
probability of a married person purchasing a bike is computed from how often a married
person and a bike buyer appear together in the training data without considering any other
columns. The probability of a new case is just the normalized product of the individual
probabilities.Several viewers display data from the fi nished model:■ (^) Dependency Network: Displays both input and predictable columns as nodes with
arrows indicating what predicts what; a simple example is shown in Figure 57-6.
Move the slider to the bottom to see only the most signifi cant predictions. Click a
node to highlight its relationships.
■ (^) Attribute Profile: This shows all variables and predictable outcomes in a single
matrix. Each cell of the matrix is a graphical representation of that variable’s dis-
tribution for a given outcome. Click a cell (chart) to see the full distribution for
that outcome/variable combination in the Mining Legend, or hover over a cell for
the same information in a tooltip.
■ Attribute Characteristics: This viewer displays the list of characteristics associ-
ated with the selected outcome.
■ Attribute Discrimination: This viewer is similar to the Characteristics Viewer, but
it shows which characteristics favor one outcome versus another.
Association Rules
The association rules algorithm operates by fi nding attributes that appear together in cases
with enough frequency to be signifi cant. These attribute groupings are called itemsets,
which are in turn used to build the rules used to generate predictions. Although you can
use Association Rules for many tasks, it is specifi cally suited to market basket analysis.
Generally, you can prepare data for market basket analysis using a nested table, whereby
c57.indd 1294c57.indd 1294 7/31/2012 10:35:03 AM7/31/2012 10:35:03 AM
http://www.it-ebooks.info