FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2125863463> ?p ?o ?g. }

• W2125863463 abstract "In simulated soccer, the level of performance depends on many interrelated subsystems, one of such systems is the analysis and subsequent exploitation of the opposing teams’ behaviour. The base code for most simulated soccer teams only partially changes from one competition to another, thus the previous competitions logs and binaries (the historical data) can be used for such analysis. Research in this field is still largely unexplored, not having proved itself in competitions like RoboCup Soccer Simulation 2D league yet. In this thesis a three step approach for such system is proposed. The first component is responsible for generating a human-readable opponent model based on detected events (match statistics), such events are defined by an expert panel and their detection is manually validated, this tool is also used to analyse the matches of the 2009 RoboCup simulation league competition. The manual validation indicates that the tool is capable of successfully detecting the desired events with minimal flaws. The second component focus on classifying an opponent based on the previously described model, to achieve those results, the opponents are clustered according to their characteristics using the K-means clustering algorithm. Then, based on the previously described model three classifiers are trained to identify to which cluster corresponds each model. The classifiers all obtain a classification rate superior to 80%–SVM (96.4%), Random Forest (93.59%) and Bagging (80.151%). The Friedman rank test rejected the null hypothesis of equivalence between the three classifiers (Bagging, Random Forest and SVM) and the Nemenyi test showed that SVM and Random Forest are better than Bagging but cannot show which is better between SVM and Random Forest. In the light of these results and considering that SVM is the fastest of the three to train, it is considered the best. The final component encompasses all of the previous components and is responsible for advising its team in a real-time scenario. To be able to do this the best tactic to face each opponent cluster is identified before the match, using that information, the SVM classifier and an online generated opponent model, it is able to identify the best tactic for each of the opponent play styles. The results are obtained by comparing the results of the FC Portugal team without and with this module at varying advise intervals. The null hypothesis is rejected by the Friedman rank test, but the Bonferroni-Dunn test does not allow to conclude if the changes made result in better or worse behaviour than the control group. In the future the obtained results can be further studied. The project can be extended by varying more tactic parameters or even applied it to another sport or environment (real instead of simulated)." @default.
• W2125863463 created "2016-06-24" @default.
• W2125863463 creator A5065962845 @default.
• W2125863463 date "2010-01-01" @default.
• W2125863463 modified "2023-09-27" @default.
• W2125863463 title "Exploiting opponent behavior in multi-agent systems" @default.
• W2125863463 cites W1482259636 @default.
• W2125863463 cites W1484799132 @default.
• W2125863463 cites W1495978126 @default.
• W2125863463 cites W1505354086 @default.
• W2125863463 cites W1521835442 @default.
• W2125863463 cites W1534818666 @default.
• W2125863463 cites W1554944419 @default.
• W2125863463 cites W1556175511 @default.
• W2125863463 cites W1564534945 @default.
• W2125863463 cites W1565746575 @default.
• W2125863463 cites W1581276390 @default.
• W2125863463 cites W1589729680 @default.
• W2125863463 cites W1608092299 @default.
• W2125863463 cites W1927792210 @default.
• W2125863463 cites W1973919178 @default.
• W2125863463 cites W1980745722 @default.
• W2125863463 cites W1982453322 @default.
• W2125863463 cites W2006410374 @default.
• W2125863463 cites W200759931 @default.
• W2125863463 cites W2011430131 @default.
• W2125863463 cites W2034587361 @default.
• W2125863463 cites W2067885219 @default.
• W2125863463 cites W2076124809 @default.
• W2125863463 cites W2076723282 @default.
• W2125863463 cites W2091751641 @default.
• W2125863463 cites W2097744910 @default.
• W2125863463 cites W2101915609 @default.
• W2125863463 cites W2102512140 @default.
• W2125863463 cites W2107263179 @default.
• W2125863463 cites W2107280071 @default.
• W2125863463 cites W2107512548 @default.
• W2125863463 cites W2114860635 @default.
• W2125863463 cites W2115288602 @default.
• W2125863463 cites W2117979647 @default.
• W2125863463 cites W2122410182 @default.
• W2125863463 cites W2145790759 @default.
• W2125863463 cites W2163453774 @default.
• W2125863463 cites W2252957903 @default.
• W2125863463 cites W2610559872 @default.
• W2125863463 cites W2736978630 @default.
• W2125863463 cites W2782037066 @default.
• W2125863463 cites W2911964244 @default.
• W2125863463 cites W2912934387 @default.
• W2125863463 cites W101836494 @default.
• W2125863463 cites W2119547536 @default.
• W2125863463 cites W2546400171 @default.
• W2125863463 cites W2624347098 @default.
• W2125863463 hasPublicationYear "2010" @default.
• W2125863463 type Work @default.
• W2125863463 sameAs 2125863463 @default.
• W2125863463 citedByCount "0" @default.
• W2125863463 crossrefType "dissertation" @default.
• W2125863463 hasAuthorship W2125863463A5065962845 @default.
• W2125863463 hasConcept C114614502 @default.
• W2125863463 hasConcept C119857082 @default.
• W2125863463 hasConcept C121332964 @default.
• W2125863463 hasConcept C12267149 @default.
• W2125863463 hasConcept C124101348 @default.
• W2125863463 hasConcept C154945302 @default.
• W2125863463 hasConcept C164226766 @default.
• W2125863463 hasConcept C168167062 @default.
• W2125863463 hasConcept C169258074 @default.
• W2125863463 hasConcept C33923547 @default.
• W2125863463 hasConcept C41008148 @default.
• W2125863463 hasConcept C73555534 @default.
• W2125863463 hasConcept C97355855 @default.
• W2125863463 hasConceptScore W2125863463C114614502 @default.
• W2125863463 hasConceptScore W2125863463C119857082 @default.
• W2125863463 hasConceptScore W2125863463C121332964 @default.
• W2125863463 hasConceptScore W2125863463C12267149 @default.
• W2125863463 hasConceptScore W2125863463C124101348 @default.
• W2125863463 hasConceptScore W2125863463C154945302 @default.
• W2125863463 hasConceptScore W2125863463C164226766 @default.
• W2125863463 hasConceptScore W2125863463C168167062 @default.
• W2125863463 hasConceptScore W2125863463C169258074 @default.
• W2125863463 hasConceptScore W2125863463C33923547 @default.
• W2125863463 hasConceptScore W2125863463C41008148 @default.
• W2125863463 hasConceptScore W2125863463C73555534 @default.
• W2125863463 hasConceptScore W2125863463C97355855 @default.
• W2125863463 hasLocation W21258634631 @default.
• W2125863463 hasOpenAccess W2125863463 @default.
• W2125863463 hasPrimaryLocation W21258634631 @default.
• W2125863463 hasRelatedWork W1148468569 @default.
• W2125863463 hasRelatedWork W1542298582 @default.
• W2125863463 hasRelatedWork W2085085514 @default.
• W2125863463 hasRelatedWork W2148972769 @default.
• W2125863463 hasRelatedWork W2152457075 @default.
• W2125863463 hasRelatedWork W2339519042 @default.
• W2125863463 hasRelatedWork W2768927604 @default.
• W2125863463 hasRelatedWork W2773651678 @default.
• W2125863463 hasRelatedWork W2888408936 @default.
• W2125863463 hasRelatedWork W2907132304 @default.
• W2125863463 hasRelatedWork W2912744730 @default.
• W2125863463 hasRelatedWork W2942313648 @default.

• next