Predictive analysis can give our business an edge. In this chapter, we will go beyond the simple descriptive statistical functionalities of MicroStrategy. We will discuss how to make predictions based on existing data in hand. We will also discuss how to automate the predictive mechanism using MicroStrategy 10.
The predictive power of a model depends on data quality, and identified patterns and trends remaining unchanged. MicroStrategy is not the ideal tool for discovering these patterns and trends; SAS, R, Stata, or EViews might be better suited to the model developing stage. The strength of MicroStrategy lies in its ability to implement and automate pre-built prediction models into our database, thus saving us time and greatly increasing productivity. MicroStrategy can integrate with third-party solutions, especially with R.
This chapter will cover the following topics:
- Integrating R with MicroStrategy
- Data mining overview
- Steps for doing data mining in MicroStrategy
- Several business applications of predictive analysis
- Automating the predictive mechanism
Predictive analysis is part of Data Mining Service, and we need to meet certain licensing requirements. With the license, we have access to native analytical functions, as well as many features provided through R integrations.
Data Mining Service is part of Intelligence Server, and we need a Developer Designer license to access it.
According to the MicroStrategy 10 documentation, there are 270 native functions built into MicroStrategy, covering basic, OLAP, mathematical comparison, financial, predictive, and statistical areas. With them we can conduct linear regression, logistic regression tree regression, decision trees, association rules models, and time series models.