ADVANCED ANALYTICS

The InforSense Advanced Analytics package is a series of tools that extend InforSense. It is designed to support and enhance the work performed by data analysts using InforSense.

The suite contains the following tools:

• Analytical Algorithms - A comprehensive collection of data mining and statistical algorithms are included to support Classification, Clustering, Association, Regression and Multi-variate analysis of data.
• Classification Studio – The InforSense Classification Studio is a guided environment to enable users to choose from “best practice” predictive modeling workflows and to combine these with the analysts' own preferences for data preparation, feature selection, modeling and evaluations. Within the Classification Studio, variants of the model can be built using different processes so that models can be easily compared side by side to see which enables the faster optimization of models.
• Clustering Studio – The InforSense Clustering Studio is a guided environment to enable users to choose from “best practice” descriptive modeling workflows and to combine these with the analysts own preferences for data preparation and modeling. Within the Clustering Studio, variants of the model can be built using different processes so that models can be easily compared side by side to see which enables the faster optimization of models.
• Interactive Decision Tree – The InforSense Interactive Decision Tree enables users to decompose the decision tree model-building algorithms. This enables the user to combine their own understanding of the business problems with the aforementioned algorithms to determine the best choice of splits for a specific application.
• Analytical Connectors – InforSense provides connectors to common statistical and analysis tools such as R, Matlab and SAS. These enable the user to incorporate an organization's own methods directly into a larger analytical process.

Following are two example applications where InforSense Advanced Analytics has been applied.

Credit Risk Scoring

InforSense Classification Studio has been used to build a series of predictive models that can be used as the analytical engines within Probability of Default (PD) calculation. Using the Classification Studio, data was collected about previous loan applications and whether the loan had been defaulted. Using the Classification Studio environment, models about which customers were likely to default were built and compared. The Classification Studio was used both to aid users in the interpretation of the models, using its diagnostic tools, as well as to compare models built using different classification approaches in order to select the best model for the applications.

Microarray Data Analysis

For Microarray data analysis, the de-facto standard for normalization methods is provided by Bio-Conductor using the R Connector. From the InforSense Advanced Analytics package, users are able to combine the Bio-Conductor methods directly in an analytical workflow that enables users to normalize and then use other methods, including the InforSense Classification and Clustering Studios, to perform analytic operations.