Data Quality on Model Validity

Requirements Clarification & Assessment

1. Model Type:

   • The model in question is a logistic regression model.
   • The logistic regression model is sensitive to the scale and distribution of input data.

2. Key Variable:

   • The model heavily relies on a single variable.
   • This variable's data includes values like 50.00, 100.00, and 40.00.

3. Data Quality Issue:

   • Some values have lost their decimal points due to an error, e.g., 100.00 recorded as 10000.
   • This error results in a significant change in the magnitude of the values.

4. Impact on Model:

   • The logistic regression model is sensitive to outliers.
   • The error introduces outliers by increasing the magnitude of some values by a factor of 100.

5. Objective:

   • Determine if the model remains valid with the erroneous data.
   • Identify strategies to rectify the model if it is invalid.