This project predicts category, priority & name of person to whom ticket can be assign. User have to enter name of the software and problem description, machine learning model will predict category, priority and name of the resolver for that given ticket.
Gathered, analyzed, and prepared the data for development.Worked with Multi-target Neural Network model for prediction and classification with python and Deployed model with Flask framework on the web platform.
Also Created Auto Rebuild ML model that can rebuild itself with new data.
Final model is trained on 50 Epochs, 0.3 drop out, 0.0001 learning rate and 10 batch size with accuracy of 95.25% in category, 88.62 in priority and 65.77% in assignment. These models are trained on 13,000 rows of data.