Practical Predictive Analytics

Time-dependent data

In many analytic techniques, all data is treated as static at the time of analysis, and these techniques are not well equipped to deal with data that changes over time. Any event that measures a change in a variable over time, such as age, or changes in attitudes via surveys can be treated as time-dependent variables, and are handled well by survival techniques. This works well within a marketing concept since the treatment of different customer segments can change, and we need a mechanism to measure the effect of interventions such as promotions and advertising that can change customer behavior over time.