**8. Stages of PAA development**

This section explains a more streamlined and contextual version of cross industry standard process for data mining (CRISP-DM). It is a neutral framework that addresses data analytics from two perspectives: application and technical. It is commonly used in predictive data analytics. As we focus on these details, it needs to be pointed out here that conducting (PDA) should never be diploid simply for the sake of expressing curiosity or flaunting one's knowledge of an existing problem-solving strategy. PDA is meant to solve problems. And in order to solve these problems, significant efforts are required to justify its application. One important component of such an exercise is the identification of a relevant management challenge. Hard questions need to be asked. What specifically is the issue? What are some of the interventions that have been made? How have the intervention outcomes improved or addressed the problem? And how have these interventions contributed in mitigating these problems. A combination of these questions will help significantly in redirecting and focusing intervention strategies.

### **8.1 Problem statement**

In this stage, the business problem that needs to be addressed should be identified. The objective can be to perform a forecast of the future needs or to establish the likelihood of occurrence of a particular defect. The resulting predictive algorithm should be one that promotes the attainment of the goals and objectives that have been identified [13]. Problem statement identification also involves the definition of performance metrics that a business needs to achieve. A plan should be devised that enables the measurement of the metrics when data are input into the algorithm.
