Best Model for Prediction

Discussion in 'Big Data and Analytics' started by _39028, Nov 23, 2018.

  1. _39028

    _39028 Member

    Aug 31, 2018
    Likes Received:
    Hi All,
    I am working on a real business problem of predicting/identifying delinquent customers from my loan portfolio.
    My data set includes both financial and demographic information about the customers and his loan and saving accounts behavior. The attributes includes numeric and non numeric variables, few list of variables as give below.

    What I would like to know is the best prediction method I should use to predict this scenario.

    1. Age
    2. Gender
    3. Nationality
    4. Age at Loan Booking Date
    5. Employer - this field contact name of the employer
    6. Employment sector - Private / Govt / Semi Govt etc.
    7. City of the customer resident
    8. Branch from where the customer booked the loan
    9. Risk Ratings
    10. Salary
    11. Salary Range
    12. Average Saving Account Balance
    13. Number of Salaries received in last 2 years
    14. Number of months the salary credit date delayed in last 2 years
    15 . Average monthly credit in saving account
    16 . Average monthly Debit in saving account
    17. Delinquent ( Customers who delayed their loan installment repayment more than 90 days )

    Thank you In Advance.,

    Mahroof NM
  2. Nishant_Singh

    Nishant_Singh Well-Known Member
    Simplilearn Support

    Aug 1, 2018
    Likes Received:
    Hi Learner,

    I suggest you to use a Decision Tree model for this prediction. However, it is not a restriction as you can use any other model as well if you feel that it will give you an even better result.

    In Data Science, we have a lot of ways to achieve every single goal and it completely depends on you to decide how to approach is based on your understanding.

    I hope that this will help you.

    Nishant Singh
  3. _51422

    _51422 New Member

    Dec 10, 2018
    Likes Received:
    Hello Nishant,
    I have already attend the community and Im new in big data . Where can I start?
  4. Dhananjay Venuri

    Dhananjay Venuri Active Member

    May 2, 2016
    Likes Received:

    While i am still learning and performing simple projects.. I could relate that the data model should be multivariate...There are many good models to begin with..If i were you. I would begin exploring the datamodel analysis with the "Class" Categories classification types. My answer would vary based on Sas tool versus R tool. However i would do comparison modelling between ivanova model versus decission tree models..Again please understand no model is wrong..However the precission in selecting the data compaisons between these models has been very amusing to me... While i hardly know in depth all the advantages or disadvantages... I found these data models to be very helpful.

Share This Page