Data Analytics of an Information System Based on a Markov Decision Process and a Partially Observable Markov Decision Process

Data Analytics of an Information System Based on a Markov Decision Process and a Partially Observable Markov Decision Process

DOI: 

https://doi.org/10.30564/jcsr.v5i1.5434

Abstract

Data analytics of an information system is conducted based on a Markov decision process (MDP) and a partially observable Markov decision process (POMDP) in this paper. Data analytics over a finite planning horizon and an infinite planning horizon for a discounted MDP is performed, respectively. Value iteration (VI), policy iteration (PI), and Q-learning are utilized in the data analytics for a discounted MDP over an infinite planning horizon to evaluate the validity of the MDP model. The optimal policy to minimize the total expected cost of states of the information system is obtained based on the MDP. In the analytics for a discounted POMDP over an infinite planning horizon of the information system, the effects of various parameters on the total expected cost of the information system are studied.

Keywords: 

Predictive modelling, Information system, MDP, POMDP, Cybersecurity, Q-learning

Comments

Popular posts from this blog

Experimental Simulation of Red Sprites in a Laboratory

Analysis of Chinese Citizens’ Perception and Its Differences of City Spirit: A Case Study of Hefei City

Bond Behavior between Recycled Concrete Containing CDW and Different Types of Steel Bars