|
In approaching a customer’s data mining or forecasting challenge, Visual Numerics conducts a thorough review of the customer’s data. Visual Numerics may find that one of its classic forecasting techniques is suitable for the given situation.
Regression Analysis, Correlation Analysis, Cluster
Analysis, Analysis of Variance, and Interpolation are classic approaches
to forecasting and predictive analysis available in the IMSL Family of
products. Other data mining and forecast techniques include Covariance
Analysis, Discriminant Analysis, and a comprehensive range of probability
distributions and random number generators, which can be used for Monte
Carlo simulation.
In determining the best technique, Visual Numerics experts may ask questions such as:
Is the data observed over time?
Is the data observed over regular intervals?
Is the data continuous or categorical?
|
|
| Regression Analysis
|
- Excellent for continuous data
- Data does not need to be viewed over time
|
Discriminant
Analysis
Analysis
of Variance (ANOVA)
Design of Experiments
Logistic Regression
General Linear Model
|
- Excellent for categorical data
- Data does not need to be viewed over time
|
Interpolation |
|
|
