Crowd-ranking: a Markov-based method for ranking alternatives

B. Vaziri (Corresponding author), S. Dabadghao, Y. Yih, T.L. Morin, M. Lehto

Onderzoeksoutput: Bijdrage aan tijdschriftTijdschriftartikelAcademicpeer review

Uittreksel

Many ranking algorithms rank a set of alternatives based on their performance in a set of pairwise comparisons. In this study, a special scenario is observed in which the objective is to rate and rank a set of groups in a traditional recruiting situation, in which the groups extend offers to the set of individuals, and the individuals will select one of their available offers. The new ranking method, Crowd-Ranking, uses collective wisdom and decision-making in conjunction with Markov chains to create competitive matches between alternatives and ultimately provide a ranking of the alternatives. First, the method is evaluated by its performance in a perfect season scenario. Next, it is applied to the case of NCAA football recruiting in the power conferences (ACC, Big Ten, Big 12, Pac 12 and SEC) in the Football Bowl Subdivision. For the Big Ten conference, the method performs significantly better than popular existing services at predicting future team performance based on past recruiting rankings. For a comprehensive national ranking of the power conferences, there is no statistically significant difference between Crowd-Ranking and the other methods.

TaalEngels
Aantal pagina's17
TijdschriftOperational Research
DOI's
StatusE-publicatie vóór gedrukte publicatie - 18 mei 2017

Vingerafdruk

Markov processes
Ranking
Decision making
Alternatives
Scenarios
Pairwise Comparisons
Subdivision
Markov chain
Decision Making
Recruiting

Trefwoorden

    Citeer dit

    @article{946be8955ebd4acda81e8164afd67972,
    title = "Crowd-ranking: a Markov-based method for ranking alternatives",
    abstract = "Many ranking algorithms rank a set of alternatives based on their performance in a set of pairwise comparisons. In this study, a special scenario is observed in which the objective is to rate and rank a set of groups in a traditional recruiting situation, in which the groups extend offers to the set of individuals, and the individuals will select one of their available offers. The new ranking method, Crowd-Ranking, uses collective wisdom and decision-making in conjunction with Markov chains to create competitive matches between alternatives and ultimately provide a ranking of the alternatives. First, the method is evaluated by its performance in a perfect season scenario. Next, it is applied to the case of NCAA football recruiting in the power conferences (ACC, Big Ten, Big 12, Pac 12 and SEC) in the Football Bowl Subdivision. For the Big Ten conference, the method performs significantly better than popular existing services at predicting future team performance based on past recruiting rankings. For a comprehensive national ranking of the power conferences, there is no statistically significant difference between Crowd-Ranking and the other methods.",
    keywords = "Crowds, Decision making, Markov chains, Ranking",
    author = "B. Vaziri and S. Dabadghao and Y. Yih and T.L. Morin and M. Lehto",
    year = "2017",
    month = "5",
    day = "18",
    doi = "10.1007/s12351-017-0324-7",
    language = "English",
    journal = "Operational research quarterly",
    issn = "0030-3623",
    publisher = "Pergamon Press Ltd.",

    }

    Crowd-ranking : a Markov-based method for ranking alternatives. / Vaziri, B. (Corresponding author); Dabadghao, S.; Yih, Y.; Morin, T.L.; Lehto, M.

    In: Operational Research, 18.05.2017.

    Onderzoeksoutput: Bijdrage aan tijdschriftTijdschriftartikelAcademicpeer review

    TY - JOUR

    T1 - Crowd-ranking

    T2 - Operational research quarterly

    AU - Vaziri,B.

    AU - Dabadghao,S.

    AU - Yih,Y.

    AU - Morin,T.L.

    AU - Lehto,M.

    PY - 2017/5/18

    Y1 - 2017/5/18

    N2 - Many ranking algorithms rank a set of alternatives based on their performance in a set of pairwise comparisons. In this study, a special scenario is observed in which the objective is to rate and rank a set of groups in a traditional recruiting situation, in which the groups extend offers to the set of individuals, and the individuals will select one of their available offers. The new ranking method, Crowd-Ranking, uses collective wisdom and decision-making in conjunction with Markov chains to create competitive matches between alternatives and ultimately provide a ranking of the alternatives. First, the method is evaluated by its performance in a perfect season scenario. Next, it is applied to the case of NCAA football recruiting in the power conferences (ACC, Big Ten, Big 12, Pac 12 and SEC) in the Football Bowl Subdivision. For the Big Ten conference, the method performs significantly better than popular existing services at predicting future team performance based on past recruiting rankings. For a comprehensive national ranking of the power conferences, there is no statistically significant difference between Crowd-Ranking and the other methods.

    AB - Many ranking algorithms rank a set of alternatives based on their performance in a set of pairwise comparisons. In this study, a special scenario is observed in which the objective is to rate and rank a set of groups in a traditional recruiting situation, in which the groups extend offers to the set of individuals, and the individuals will select one of their available offers. The new ranking method, Crowd-Ranking, uses collective wisdom and decision-making in conjunction with Markov chains to create competitive matches between alternatives and ultimately provide a ranking of the alternatives. First, the method is evaluated by its performance in a perfect season scenario. Next, it is applied to the case of NCAA football recruiting in the power conferences (ACC, Big Ten, Big 12, Pac 12 and SEC) in the Football Bowl Subdivision. For the Big Ten conference, the method performs significantly better than popular existing services at predicting future team performance based on past recruiting rankings. For a comprehensive national ranking of the power conferences, there is no statistically significant difference between Crowd-Ranking and the other methods.

    KW - Crowds

    KW - Decision making

    KW - Markov chains

    KW - Ranking

    UR - http://www.scopus.com/inward/record.url?scp=85019619771&partnerID=8YFLogxK

    U2 - 10.1007/s12351-017-0324-7

    DO - 10.1007/s12351-017-0324-7

    M3 - Article

    JO - Operational research quarterly

    JF - Operational research quarterly

    SN - 0030-3623

    ER -