Link to GitHub repository - https://github.com/ajayrao1983/TitanicReinforcmentLearningExample
##Libraries Used: Python:
- numpy
- pandas
- collections
- torch
- os
##Folder Structure and Files in Repository .
|--Data
| |- train.csv # Training Set
| |- test.csv # Test Set
|- TitanicEnv.py # Defining the Titanic Environment Class
|- AgentTrainer.py # Defining and training Agent Class
|- usingTrainedAgent.py # Calling the trained agent and processing the test data set
|--README.md
##Project Description
This is an attempt at using Reinforcement Learning for unlabelled data classification tasks, assuming there is a way to:
- Model the agent
- Model the environment
- Evaluate or score the actions of the agent
The resulting model scored a 78% accuracy on the test dataset.