WBS Analytics is providing data analytics consultancy services to several industries. One of their projects is for an automobile insurance company. The data science team plans to develop a data mining system to predict whether an insurance claim is legitimate or not. Some of the customers of this automobile company provide false information to the company in order to get more payment which leads to greater financial losses. The data science team builds two data mining models for this project. The company provided historical claim data for the data mining model. The available dataset has 20,000 records and there 15,000 legitimate claims and 5,000 not legitimate (default) claims. The data science team divides the available data into training and test sets. Training dataset has 10,000 records and it is directly used to build the predictive models. The test data has remaining 10,000 records and there are 8,000 legitimate claims and 2,000 default claims. The following results are obtained for both models during the model evaluation. Model I: Model I predicted 8,500 claims as legitimate and 7,500 out of 8,500 actually correspond to the legitimate claims and hence, predicted correctly. The remaining 1,000 records actually belong to the not legitimate (default) claims. The model predicted 1,500 records as not legitimate and 1000 out of 1,500 actually correspond to the not legitimate claims and hence, predicted correctly. The remaining 500 records were actually legitimate claims. Model II: Model II predicted 7,600 claims as legitimate and 6,800 out of 7,600 actually correspond to the legitimate claims and hence, predicted correctly. The remaining 800 records actually belong to the not legitimate (default) claims. The model predicted 2,400 records as not legitimate and 1,200 out of 2,400 actually correspond to the not legitimate claims and hence, predicted correctly. The remaining records were actually legitimate claims.