Developed an inverse reinforcement learning algorithm with linear decomposition of the environment to find out the skill sets that act as the basis functions of animal behaviors. Used the trajectory data and recover the reward function efficiently and locate a series of skill sets.