These features analyze user data to extract behavioral insights, providing a deeper understanding of user habits beyond just financial data. The features are:
Feature name | Definition |
---|---|
Buys_insurance | Whether or not an individual buys insurance, in a period of 0-90 days, returns True or False |
Count_gratitude_events_{t0}_{t1} | The number of SMS received indicating a show of gratitude. The counts are summed over a period of t0 to t1 days history prior to the prediction date, where the time windows are 0-30, 31-90, or 0-90 days |
Count_loan_seeking_events{t0}_{t1} | The number of SMS received indicating a loan application not completed or completed but still not processed. The counts are summed over a period of t0 to t1 days history prior to the prediction date, where the time windows are 0-10, 0-30, 31-90, or 0-90 days |
Count_school_fee_payment_reminder_events{t0}_{t1} | The number of SMS received reminding a parent to pay school fees or that they have a school fee balance. The counts are summed over a period of t0 to t1 days history prior to the prediction date, where the time windows are 0-30, 31-90, or 0-90 days |
Gender | Returns string Male Or Female based on sms or None if cannot be determined using the sms |
Owns_car | Whether or not an individual owns car, in a period of 0-90 days, returns True or False |
Pays_school_fees | Whether or not an individual pays school fees, in a period of 0-90 days, returns True or False |
Teaches_at_school | Whether or not an individual is a teacher, determined when addressed as teacher, in a period of 0-90 days, returns True or False |
Works_in_agriculture | Whether or not an individual is a farmer, determined when addressed as farmer, in a period of 0-90 days, returns True or False |