Font Size: a A A

The Developmental Of Recruiting Win-stay & Loss-switch Strategies In Two-armed Bandit Problem

Posted on:2018-12-20Degree:MasterType:Thesis
Country:ChinaCandidate:Y X LiuFull Text:PDF
GTID:2335330542963487Subject:Applied Psychology
Abstract/Summary:PDF Full Text Request
In our daily life,In daily life,we are always faced with a variety of choices involving probability learning,we will be based on the results of the feedback to stay or adjust our behavior.The studies of adults show that,they will stay after “a win trial”(Win-stay)but switch after “a loss trial”(Loss-switch).However,the tendency of this behavior,or whether the use of learning strategies is influenced by the environment,the probability of the option,and it is unknown that how the influences changes development in the process of cognitive.In order to investigate these issues,in 5 age groups(children 4-6 years of age,children 7-9 years of age,10-12 years old children,adults and the elderly)were tested by the two-armed bandit problem(bubble game),subjects were asked to as much as possible to award points.Based on the analysis of the ratio of Switch under different conditions,Study 1 and Study 2 draw the following conclusions:1.Overall,the good or no good of the options,the current winning or losing options and the win or loss environmental task have an impact on Win-stay and Loss-switch strategies.It will change with development,and the main developmental change is in the win environmental task(the probability of winning the overall higher than 0.5).2.In addition to the old man,the age of the subjects in win environmental task(the overall probability of winning less than 0.5),there was no significant difference between the various conditions.reaction tendency showed that,more inclined to choose update-Switch.3.From the results of adults,in win environmental task,the subjects in both the good option(0.7)or no good option(0.5),are clearly showing the use of Win-stay and Loss-switch strategies.And Switch is significantly higher than the good option in the no good option.4.10-12 years of age is a turning point in the development of children.At this time,in win environmental task is the same as adults: in the good or no good options,the ratio of Switch after winning or losing feedback is different.5.Although 4-6 years old children prefer to Switch under all conditions,but in the good option and have a positive feedback,the proportion of Switch is significantly lower than other conditions.This shows that children have begun to use the Win-stay strategy,but only in the case of a higher proportion of options win.6.7-9 years children ues Win-stay strategy to expand the proportion of options which gains lower,then both good or no good options,the subjects in the feedback of winning after the selection of the ratio of Switch decreased significantly after all with respect to the output.This shows that children can apply Win-stay strategy to lower probability options.7.The elderly adults' behavioral results especially unique characteristics,from the result of this research,the elderly adults did not use the optimization strategy(Win-stay,Loss-switch)which used by adults,but but show strategies of Win-switch and Loss-stay,and in loss environmental task is more obvious.We believe that maybe the elderly prefer thinking that the rules are changing at any time,and now the option to bring rewards may soon be no longer rewarded.Generally speaking,Win-stay is not a simple behavioral tendency based on feedback,and its appearance depends on the estimation of the overall probability of the selected option.For young children,Switch seems to be a default response,they maintained a high level of conversion(0.7),after the switch level only in the high probability advantage of the option to get positive feedback,obviously decreased compared with the other conditions.For older children,the performance was similar to that of adults,and Win-stay and Loss-switch tended to be more dominant in both the dominant and inferior options,and the ratio of Switch was different under different conditions.The old people showed the Win-switch and Loss-stay response strategies,and were more obvious in the environment.To sum up,Win-stay and Loss-switch are a kind of "use" strategy,which is used only when the overall environment is positive.
Keywords/Search Tags:two-armed bandit problem, feedback, Win-stay, Loss-switch strategies, children
PDF Full Text Request
Related items