• This was the multiclass classification problem to find out the reason for mis paid amount paid to the patients.
• Build the 2-stage model.
• Level 1 model- To find the 2 main reasons for mis paid claims.
• Level 2 model- To find the sub reason for the main reasons.
• Build and deployed the model, this model helped to find out the areas which are prone to error and need focus on.

• Cleaned the data using Regular Expression, Removed Stop words from text, Converted Categorical values to numerical using dummy encoder.
• Used stemming to convert data to base form.
• I used the Naïve Bayes Classifier model and got the accuracy score about 0.98.
• Techniques - NLP, Python.

Cleaned the data using Regular Expression, Removed Stop words from text, ConvertedCategorical values to numerical using dummy encoder.Used stemming to convert data to base form.I used the Naïve Bayes Classifier model and got the accuracy score about 0.98.