In this chapter, we will cover batch analytics using Hadoop. Traditional marketing campaigns which involve direct mailing to the entire customer base are an expensive project for companies. A marketer does not know in advance who is going to respond to a marketing campaign. Generally, the response to a campaign is considered in the form of an action that we expect the recipient to take as the outcome of a campaign. If a campaign fails to evoke the expected or desired response then it is not considered a successful campaign.
In this chapter, we will use Hadoop to perform batch analytics on a customer database to increase the likelihood of customer response in a marketing campaign. We will follow the following steps towards building the solution:
Understanding of classification as a supervised learning method
Building a machine learning model using historical response data
Using the machine learning model in a MapReduce job to generate a list of customers...