6781097116

6781097116



ISBN: 978-989-95618-4-7

For post-flop rounds. the Expectation Maximization algorithm resulted in five different clusters. The centroids of the clusters are as follows.

Tablc 4. Post-Flop centroids

Feature

Cluster #0

13%

Cluster #1

15%

Cluster #2 14%

Cluster #3 17%

Cluster #4 42%

\vin_prob

0.6946

0.5297

0.7052

0.6463

0.5306

position

Late

Late

Early/Late

Early

Earh

possible eamings

0.1118

0.0867

0.6305

0.1133

0.1186

action

Raise

Cali

Call/Raise

Call/Raise

Cali

betted money

0.0574

0.0008

0.3586

0.0558

0

The most common actions after the flop (Clustcr #4 and 42% of the plays) are Checks (Calls without any money involved).

7 Clustering player types

After defining common move types both in Pre-Flop and Post-Flop rounds. clustering algorithms were used to determine common strategies. To do this. all players' actions on the database were considered.

Initially a cluster was assigned to each action. To assign a clustcr we calculated the Euclidean distance (Eąuation 4) between the action and each cluster. The assigned cluster is the one with the lowest distance.

d(p, <l) = VS"=i(9i ~ Pi)2    (•*)

Therefore. for each player the freąuency of actions of each cluster was calculated. The following table describes the file structure ARFF rcpresentathe of the players' strategies.

Table 5. Training set characteristics to determine player strategies.

Name

Type

Description

Name

String

The name of the player.

Pre_C0

Numeric [0.1]

Relatwe freąuency of actions in the type of cluster 0.

Pre-Flop of

Pre_Cl

Numeric [0.1]

Relative freąuency of actions in the type of cluster 1.

Pre-Flop of

Pre_C2

Numeric [0.1]

Relatwe freąuency of actions in the type of cluster 2.

Pre-Flop of

Pre_C3

Numeric [0.1]

Relatwe freąuency of actions in the type of cluster 3.

Pre-Flop of

Pre_C4

Numeric [0.1]

Relatwe freąuency of actions in the type of cluster 4.

Pre-Flop of

Pre_C5

Numeric [0.1]

Relatiye freąuency of actions in the type of cluster 5.

Pre-Flop of

Post_C0

Numeric [0.1]

Relatwe freąuency of actions in the type of cluster 0.

Post-Flop of



Wyszukiwarka

Podobne podstrony:
EPIA 2011 ISBN: 978-989-95618-4-7 no information about the cards of the player when they folded the
EPIA 2011 ISBN: 978-989-95618-4-7Identifying Player’s Strategies in No Limit Texas Hold’em Poker thr
ISBN: 978-989-95618-4-7 Regarding possible eamings. it is possible to obsener that the stakes are mu
EPIA 2011 ISBN: 978-989-95618-4-7 Since the freąuencies are relatwe. the sum of the freąuencies of
EPIA 2011 ISBN: 978-989-95618-4-7 agents to become better players. Tliis is because when the agent i
EPIA 2011 ISBN: 978-989-95618-4-7 Thus. the purpose of tliis study is to identify groups of similar
EPIA 2011GC ®_© O © Fig. 1. Poker table initial configuration ©© ISBN: 978-989-95618-4-7 After
EPIA 2011 ISBN: 978-989-95618-4-7 If the player folds 72% or inore hands then he is considered to be
EPIA 2011 ISBN: 978-989-95618-4-7 once and then evaluates them. Some well-known partitioning methods
EPIA 2011 ISBN: 978-989-95618-4-7 By measuring the freąuency of each gamę move by the extracted play
EPIA 2011 ISBN: 978-989-95618-4-7 *** SUMMARY *** Total Pot($12) Seat 4: PtgusfQqsttogld64pQOGw (big
EPIA 2011 ISBN: 978-989-95618-4-7 To extract the gamę State of a database of played poker, an applic
EPIA 2011 ISBN: 978-989-95618-4-7 In order to use Weka. the extracted gamę States must be converted
File0008 Hi! I m Martin l m 17 years old and Tm looking for e-mail friends of the same age from
ISBN 978-83-61756-10-1 1. Eufemia (antę 1299-1359) 2. Peregryn z Opola (ca 1260-post 1333) 3. Wincen
Bibliogr.s. 453-468. Indeks. ISBN 978-83-7030-824-7 1. Duchowieństwo katolickie - Niemcy - 18-20 w.
Tyt. oryg.: How to cope with people who dnve you crazy 1998. ISBN 978-83-7660-448-0 : zl 19,90 1. St
ISBN 978-83-7252-582-6 1. Mniejszości religijne - Italska - od 1944 r. 316.347(438) b28057867
Tekst częśc. tl. Streszcz. ang. przy pracach. ISBN 978-83-7780-413-1 1. Organizacje międzynarodowe 2

więcej podobnych podstron