Promoting Neutral Net Model to Production

DAT 690 – M.6 – Journal: Midway Reflection
6-2 Journal Status Report
Assignment
Student Name
Course ID
Instructor
Institutional Affiliation
Midway Reflection
The single design that I selected for my loan default risk probability model is the neural net model. The model can be used to determine the credit default risk for a particular loan applicant by calculating risk based on certain attributes that must be entered into the model. What is working in the neural net model is that it can accept data inputs into the predictor variables of an applicant. Secondly, it can tentatively classify an applicant based on specific predictor attributes. However, what is not working in the model is reprocessing a new target variable once customer attributes have been captured. Secondly, the overall model still needs customization to accept numerous predictor variables of a loan applicant. Thirdly, the model is yet to classify applicants on a binary scale accurately. Work is in progress to ensure that the binary scale is operational and effective in determining the customer’s default probability status.
I plan to promote the neural net model to production. Unlike the conventional models such as Naïve Bayes and Logistical regression that calculate an applicant’s default probability based on percentage scores, it classifies default probability on a binary scale. Besides, it is preferred to the conve...