Creating a dataset report

The first step in creating a predictive model is to develop a dataset report in MicroStrategy. It is recommended that you use MicroStrategy as the data mining source for the following reasons:

Enterprise data warehouse: MicroStrategy projects typically use the enterprise data warehouse, which often contains clean, high-quality data. Since creating a good dataset report is typically 50-75% of the data mining effort, using MicroStrategy can greatly reduce that effort.
Analytical functions: It is easy to create new metrics using MicroStrategy’s vast library of over 200 analytical functions, which range from OLAP functions such as Rank, Percentile, and Moving Averages to advanced statistical functions. Raw facts in the database are easily turned into simple, and sometimes sophisticated, predictive inputs using these functions.
Relational databases: MicroStrategy is optimized for all the major relational database vendors, which means that dataset reports are optimized to take advantage of these database capabilities.
Security model: MicroStrategy’s robust security model ensures that users can only access the data that they are permitted to access. The MicroStrategy business intelligence platform can address privacy-preserving issues as the predictive model is developed and also when it is deployed.
Easy handling of reports: The dataset report can be easily created, refreshed, and accessed, even if it is large or contains complex calculations. Users do not have to be database administrators nor do they have to code queries to access the data they need.
Consistent use of variable definitions: The exact definitions of variables that were used to develop the model are re-used when the predictive model is deployed. This process is performed automatically, which ensures the results are consistent with the way the model was developed.
Data sampling: The dataset report can be sampled from a large number of records.

For further information, see:

Data mining dataset reports
Guidelines for creating a dataset report
Inputs for predictive metrics
Using non-MicroStrategy dataset reports
Did you find this helpful?
Waren Ihnen die Informationen hilfreich?
Ces informations vous ont-elles été utiles ?
Questo contenuto ti è stato utile?
¿Le ha resultado útil?
Você achou isso útil?
도움이 되었습니까?