Customer Value Optimization
Help service providers understand their customers to serve them better, thus Analytics is a step ahead of business intelligence (BI).
Classification mainly deals with arranging the data into multiple groups by aiding discovery of a predictive learning function that classifies a data item into one of several possible classes. Models are generated using a training set to predict categorical class labels and then utilized to classify unseen data in a testing set. Classification can be done either on grouped datasets with a set of observations with the objective of correctly establishing the grouping of data into different possible classes. Alternatively, we may know for certain that there is a set of many classes and the objective is to establish a rule whereby we can classify observations into one of the existing classes. The former is called Unsupervised learning (or clustering) and the latter is called Supervised learning.
Methods: Linear/Non-Linear Regression, Neural Networks, Decision Trees, Support Vector Machines (SVM), k-NN, Self Organizing Maps (SOM), Fuzzy Logic.
- Classifying credit applicants as low, medium or high risk;
- Customer profiling and segmentation into different purchase behavior segments based on physiographic, behavioral and demographical variables.