user manual

37
Understanding Data Mining
The Self-Learning Resp onse Model (SLRM) node enables you to build a model in
which a single new case, or small number of new cases, can be used to reestimate the
model without having to retrain the model using all data.
The Time Series node estimates exponential smoot hing, univariate Autor egressive
Integrated Moving Average (ARIMA), and multivariate ARIMA (or transfer function)
models for time series data a nd produces forecasts of future performance. A Time
Series node must always b e preceded by a Time Intervals node.
The k-Nearest Neighbor (KNN) node associates a new case with the category or value
of the k objects nearest to it in the predictor space, where k is an integer. Similar cases
are near each other and dissimilar cases are distant from each other.
Association Models
Association models find patterns in your data where one or more entities (such as even ts ,
purchases, or attributes) a r e associated with one or more other entities. The models construct rule
sets that define these relationsh ips. Here the fields within the data can act as b oth inputs and
targets. You co uld find these associations manua lly, but association ru le algorithms do so much
more quickly, and can explore more complex patterns. Apriori and Carma models are examples of
the use of such algorithms. One other type o f association mod el is a sequence detection model,
which finds sequential patterns in time-st r uctured data.
Associa tion models are most useful when predicting multiple outcomes—for example , customers
who bought pr oduct X also bought Y and Z. Associatio n models associa te a particular conclusion
(such as the de cision to buy something) with a set of conditions. The advantage of ass ociation rule
algorithms over the more standard decision tree algorithms (C5.0 and C&RT) is that associa tions
can exist between any of the attributes. A decision tree alg orithm will build r ules with on ly a
single conclusion, whereas association alg orithms attempt to find many rules, each of which may
have a different conclusion.
Association nodes
The Apriori node extracts a set of rules from the data, pulling out the rules with
the hig hest information content. Apriori offers five different methods of selecting
rules and uses a sophisticated indexing s cheme to process large data set s efficiently.
For large problems, Apriori is generally faster to train; it has no arb i trary limit on
the number of rules th at can be retained, and it can handle rules with up to 32