Feature analysis for modeling game content quality

Noor Shaker, Georgios N. Yannakakis, Julian Togelius

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Abstract

    One promising avenue towards increasing player entertainment for individual game players is to tailor player experience in real-time via automatic game content generation. Modeling the relationship between game content and player preferences or affective states is an important step towards this type of game personalization. In this paper we analyse the relationship between level design parameters of platform games and player experience. We introduce a method to extract the most useful information about game content from short game sessions by investigating the size of game session that yields the highest accuracy in predicting players' preferences, and by defining the smallest game session size for which the model can still predict reported emotion with acceptable accuracy. Neuroevolutionary preference learning is used to approximate the function from game content to reported emotional preferences. The experiments are based on a modified version of the classic Super Mario Bros game. We investigate two types of features extracted from game levels; statistical level design parameters and extracted frequent sequences of level elements. Results indicate that decreasing the size of the feature window lowers prediction accuracy, and that the models built on selected features derived from the whole set of extracted features (combining the two types of features) outperforms other models constructed on partial information about game content.

    Original languageEnglish (US)
    Title of host publication2011 IEEE Conference on Computational Intelligence and Games, CIG 2011
    Pages126-133
    Number of pages8
    DOIs
    StatePublished - 2011
    Event2011 7th IEEE International Conference on Computational Intelligence and Games, CIG 2011 - Seoul, Korea, Republic of
    Duration: Aug 31 2011Sep 3 2011

    Other

    Other2011 7th IEEE International Conference on Computational Intelligence and Games, CIG 2011
    CountryKorea, Republic of
    CitySeoul
    Period8/31/119/3/11

    Fingerprint

    Experiments

    ASJC Scopus subject areas

    • Artificial Intelligence
    • Computer Vision and Pattern Recognition
    • Human-Computer Interaction
    • Software

    Cite this

    Shaker, N., Yannakakis, G. N., & Togelius, J. (2011). Feature analysis for modeling game content quality. In 2011 IEEE Conference on Computational Intelligence and Games, CIG 2011 (pp. 126-133). [6031998] https://doi.org/10.1109/CIG.2011.6031998

    Feature analysis for modeling game content quality. / Shaker, Noor; Yannakakis, Georgios N.; Togelius, Julian.

    2011 IEEE Conference on Computational Intelligence and Games, CIG 2011. 2011. p. 126-133 6031998.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Shaker, N, Yannakakis, GN & Togelius, J 2011, Feature analysis for modeling game content quality. in 2011 IEEE Conference on Computational Intelligence and Games, CIG 2011., 6031998, pp. 126-133, 2011 7th IEEE International Conference on Computational Intelligence and Games, CIG 2011, Seoul, Korea, Republic of, 8/31/11. https://doi.org/10.1109/CIG.2011.6031998
    Shaker N, Yannakakis GN, Togelius J. Feature analysis for modeling game content quality. In 2011 IEEE Conference on Computational Intelligence and Games, CIG 2011. 2011. p. 126-133. 6031998 https://doi.org/10.1109/CIG.2011.6031998
    Shaker, Noor ; Yannakakis, Georgios N. ; Togelius, Julian. / Feature analysis for modeling game content quality. 2011 IEEE Conference on Computational Intelligence and Games, CIG 2011. 2011. pp. 126-133
    @inproceedings{5f75f2e1a482484298029c0694c3aca2,
    title = "Feature analysis for modeling game content quality",
    abstract = "One promising avenue towards increasing player entertainment for individual game players is to tailor player experience in real-time via automatic game content generation. Modeling the relationship between game content and player preferences or affective states is an important step towards this type of game personalization. In this paper we analyse the relationship between level design parameters of platform games and player experience. We introduce a method to extract the most useful information about game content from short game sessions by investigating the size of game session that yields the highest accuracy in predicting players' preferences, and by defining the smallest game session size for which the model can still predict reported emotion with acceptable accuracy. Neuroevolutionary preference learning is used to approximate the function from game content to reported emotional preferences. The experiments are based on a modified version of the classic Super Mario Bros game. We investigate two types of features extracted from game levels; statistical level design parameters and extracted frequent sequences of level elements. Results indicate that decreasing the size of the feature window lowers prediction accuracy, and that the models built on selected features derived from the whole set of extracted features (combining the two types of features) outperforms other models constructed on partial information about game content.",
    author = "Noor Shaker and Yannakakis, {Georgios N.} and Julian Togelius",
    year = "2011",
    doi = "10.1109/CIG.2011.6031998",
    language = "English (US)",
    isbn = "9781457700095",
    pages = "126--133",
    booktitle = "2011 IEEE Conference on Computational Intelligence and Games, CIG 2011",

    }

    TY - GEN

    T1 - Feature analysis for modeling game content quality

    AU - Shaker, Noor

    AU - Yannakakis, Georgios N.

    AU - Togelius, Julian

    PY - 2011

    Y1 - 2011

    N2 - One promising avenue towards increasing player entertainment for individual game players is to tailor player experience in real-time via automatic game content generation. Modeling the relationship between game content and player preferences or affective states is an important step towards this type of game personalization. In this paper we analyse the relationship between level design parameters of platform games and player experience. We introduce a method to extract the most useful information about game content from short game sessions by investigating the size of game session that yields the highest accuracy in predicting players' preferences, and by defining the smallest game session size for which the model can still predict reported emotion with acceptable accuracy. Neuroevolutionary preference learning is used to approximate the function from game content to reported emotional preferences. The experiments are based on a modified version of the classic Super Mario Bros game. We investigate two types of features extracted from game levels; statistical level design parameters and extracted frequent sequences of level elements. Results indicate that decreasing the size of the feature window lowers prediction accuracy, and that the models built on selected features derived from the whole set of extracted features (combining the two types of features) outperforms other models constructed on partial information about game content.

    AB - One promising avenue towards increasing player entertainment for individual game players is to tailor player experience in real-time via automatic game content generation. Modeling the relationship between game content and player preferences or affective states is an important step towards this type of game personalization. In this paper we analyse the relationship between level design parameters of platform games and player experience. We introduce a method to extract the most useful information about game content from short game sessions by investigating the size of game session that yields the highest accuracy in predicting players' preferences, and by defining the smallest game session size for which the model can still predict reported emotion with acceptable accuracy. Neuroevolutionary preference learning is used to approximate the function from game content to reported emotional preferences. The experiments are based on a modified version of the classic Super Mario Bros game. We investigate two types of features extracted from game levels; statistical level design parameters and extracted frequent sequences of level elements. Results indicate that decreasing the size of the feature window lowers prediction accuracy, and that the models built on selected features derived from the whole set of extracted features (combining the two types of features) outperforms other models constructed on partial information about game content.

    UR - http://www.scopus.com/inward/record.url?scp=80054014190&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=80054014190&partnerID=8YFLogxK

    U2 - 10.1109/CIG.2011.6031998

    DO - 10.1109/CIG.2011.6031998

    M3 - Conference contribution

    AN - SCOPUS:80054014190

    SN - 9781457700095

    SP - 126

    EP - 133

    BT - 2011 IEEE Conference on Computational Intelligence and Games, CIG 2011

    ER -