Finite-Sample Analysis of LSTD - Archive ouverte HAL Access content directly
Conference Papers Year : 2010

## Finite-Sample Analysis of LSTD

(1) , (1) , (1)
1
Alessandro Lazaric
• Function : Author
• PersonId : 868946
Remi Munos
• Function : Author
• PersonId : 836863

#### Abstract

In this paper we consider the problem of policy evaluation in reinforcement learning, i.e., learning the value function of a fixed policy, using the least-squares temporal-difference (LSTD) learning algorithm. We report a finite-sample analysis of LSTD. We first derive a bound on the performance of the LSTD solution evaluated at the states generated by the Markov chain and used by the algorithm to learn an estimate of the value function. This result is general in the sense that no assumption is made on the existence of a stationary distribution for the Markov chain. We then derive generalization bounds in the case when the Markov chain possesses a stationary distribution and is $\beta$-mixing.

#### Domains

Cognitive science Computer science

### Dates and versions

inria-00482189 , version 1 (09-05-2010)

### Identifiers

• HAL Id : inria-00482189 , version 1

### Cite

Alessandro Lazaric, Mohammad Ghavamzadeh, Remi Munos. Finite-Sample Analysis of LSTD. ICML - 27th International Conference on Machine Learning, Jun 2010, Haifa, Israel. pp.615-622. ⟨inria-00482189⟩

### Export

BibTeX TEI Dublin Core DC Terms EndNote Datacite

459 View