DataEngine PlugIns:
FeatureSelector
Did you ever stumble across one of these situations?
- Hundreds of attributes describing your customers and you didn't know which are the ones that determine if a customer is a "good" one or a "bad" one.
- Countless measured signals at your process and you had to find out about their influence on the current process conditions.
- A huge amount of frequencies in an acoustic signal are determined and you cannot decide which are the most relevant ones.
If your answer is "yes", your problem is feature selection: the selection of the most useful information in your data.
The dilemma is:
Solving this problem manually takes a long time while leaving
this step out decreases model accuracy and increases calculation time
tremendously.
Here is the solution that saves you time
and money:
The DataEngine PlugIn FeatureSelector automatically selects the most relevant
set of features for your application.
Essential capacities:
FeatureSelector
- is a DataEngine PlugIn for fully automatic selection of those features leading to the highest possible reclassification rate.
- provides non experts with a universal tool for solving classification problems.
- works effectively and efficiently with any number of features between 2 and 1000.
- requires the input of only a few parameters to initialize and control the selection algorithm.
- is a universal link to DataEngine's Fuzzy Clustering, Multilayer Perceptron, Kohonen and Fuzzy Kohonen Network techniques.
- has been proven in many applications including quality control, acoustic analysis and database marketing.