We have dealt with some interesting applications of machine learning in the e-commerce domain in the last couple of chapters. For the next two chapters, our big challenge will be in the financial domain. We will be using data analysis and machine learning techniques to analyze financial data from a German bank. This data will contain a lot of information regarding customers of that bank. We will be analyzing that data in two stages which include descriptive and predictive analytics.
- Descriptive: Here we will look closely at the data and its various attributes. We will perform descriptive analysis and visualizations to see the kind of features we are dealing with and how they might be related to credit risk. The data we will be dealing with here consists of labeled data already and we will be able to see how many customers were credit risks and how many weren't. We will also look closely at each feature in the data and understand its significance which will be useful in the next step.
- Predictive: Here we will focus more on the machine learning algorithms used in predictive modeling to build predictive models using the data we have already acquired in the previous step. We will be using various machine learning algorithms and testing the accuracy of the models when predicting if a customer could be a potential credit risk. We will be using labeled data to train the model and then test the models on several data instances, comparing our predicted result with the actual result to see how well our models perform.
The significance of predicting credit risks is quite useful for financial organizations, such as banks that have to often deal with loan applications from their customers. They have to then make the decision to approve or deny the loan based on information they have about the customer. If they have a robust machine learning system built in place which can analyze the data about the customer and say which customers might be credit risks, then they can prevent losses to their business by not approving loans to such customers.