Presentation is loading. Please wait.

Presentation is loading. Please wait.

Adapting Recommendation Diversity to Openness to Experience: A Study of Human Behaviour Nava Tintarev, Matt Dennis and Judith Masthoff University of Aberdeen.

Similar presentations


Presentation on theme: "Adapting Recommendation Diversity to Openness to Experience: A Study of Human Behaviour Nava Tintarev, Matt Dennis and Judith Masthoff University of Aberdeen."— Presentation transcript:

1 Adapting Recommendation Diversity to Openness to Experience: A Study of Human Behaviour Nava Tintarev, Matt Dennis and Judith Masthoff University of Aberdeen

2 Outline  Personality and recommender systems  Experiment – openness to experience and diversity  Results  Limitations  Implications for recommender systems design UMAP’2013. Rome, Italy 2

3 Personality traits  Generally it is assumed that:  a) traits are relatively stable over time,  b) traits differ among individuals (for instance, some people like to try new things while others prefer to stick to known options), and  c) traits influence behaviour (e.g. ordering familiar food at a restaurant).  Five factor model (Big Five): Extraversion, Agreeableness, Conscientiousness, Emotional Stability, and Openness to Experience  Openness to Experience: active imagination, aesthetic sensitivity, attentiveness to inner feelings, preference for variety, and intellectual curiosity UMAP’2013. Rome, Italy 3

4 Personality + Recsys == TRUE?  Recommendation is not all about accuracy.  The tailoring of recommender systems to personality has been found to  improve accuracy for sparse data sets and new users (Hu and Pu 2011),  to help predict choices for presidential candidates (Nunes 2008),  to positively impact the acceptance of a system and recommendations (Hu and Pu 2009, Wu et al 2013)  Openness to experience may make users more receptive to more diverse and potentially serendipitous recommendations. UMAP’2013. Rome, Italy 4

5 So what makes for “good” diversity UMAP’2013. Rome, Italy 5

6 Diversity != Serendipity  Unexpected and helpful (Ge et al 2010)  Topic diversification approach based on taxonomy-based dissimilarity (Ziegler et al.,2005). Impacted accuracy negatively.  Re-rank a list of top items was found to improve diversity without a great loss in accuracy (Adomavicius and Kwon, 2011)  Users preferred recommendations from a diversified set of clusters (categorical diversity?), rather than within clusters (thematic diversity?). (Abbassi et al., 2012) UMAP’2013. Rome, Italy 6

7 Research Questions  1) there may be a difference in preference for the degree of diversity in recommendations among users,  and  2) within category vs across category diversity in recommendations has not received a great deal of weight in previous literature and would benefit from empirical testing with users. UMAP’2013. Rome, Italy 7

8 Experiment: User-as-wizard UMAP’2013. Rome, Italy 8

9 Experiment - participants  Amazon Mechanical Turk  120 participants (128 excluded)  57% female, 41% male, 2% undisclosed  Openness to Experience within range for the normal population (TIPI).  Average completion time 5 minutes (up to 30) UMAP’2013. Rome, Italy 9

10 Meet Oliver (Dennis et al 2010) openness_low Oliver is not interested in abstract ideas, as he has difficulty understanding them. He does not like art, and dislikes going to art galleries. He avoids philosophical discussions. He tends to vote for conservative political candidates. He does not like poetry and rarely looks for a deeper meaning in things. He believes that too much tax money goes to supporting artists. He is not interested in theoretical discussions. Oliver is quite a nice person, and tends to enjoy talking with people. openness_high Oliver believes in the importance of art and has a vivid imagination. He tends to vote for liberal political candidates. He enjoys hearing new ideas and thinking about things. He enjoys wild flights of fantasy, getting excited by new ideas. baseline(no story) UMAP’2013. Rome, Italy 10

11 Procedure  Recommend three items (books)  Vary along three dimensions  author (0,1) Same or different  genre (0,0.3, 1) Same, similar or different  themes (0,0.3,1) Almost all themes in common, some themes in common, or no themes in common  Had to justify their choice before moving on to the next recommendation. UMAP’2013. Rome, Italy 11

12 UMAP’2013. Rome, Italy 12

13 Results UMAP’2013. Rome, Italy 13

14 Results  No statistically significant effect of story  Effect of order in sequence  But a tendency toward a difference in application of thematic and categorical diversity UMAP’2013. Rome, Italy 14

15 Effect of story ConditionDiversityAvgDiversityMax baseline1.42 (0.31)2.00 (0.42) openness_low1.41 (0.38)2.08 (0.51) openness_high 1.46 (0.30)2.14 (0.44) overall1.44 (0.33)2.08 (0.46) 2-3 things changed! Slightly higher for openness_high Difference not reliable – possible ceiling effect No correlation between the participants (aggregated TIPI score on) openness to experience and diversity. UMAP’2013. Rome, Italy 15

16 Order ConditionBook1Book2Book3 baseline1.04 (0.58)1.56 (0.49)1.67 (0.67) openness_low1.18 (0.68)1.35 (0.66) 1.72 (0.74) openness_high1.33 (0.63)1.38 (0.61)1.68 (0.72) overall1.18 (0.64)1.43 (0.60)1.69 (0.71) Book2 > div Book2 Book3 > div Book3 p < 0.01 (Bonferroni corrected) Openness_low starts lowest but `catches up’ by Book3! UMAP’2013. Rome, Italy 16

17 Ways of applying diversity ConditionAuthorGenreTheme baseline1.92 (0.70)1.22 (0.70)1.12 (0.69) openness_low1.83 (0.71)1.17 (0.63) 1.25 (0.78) openness_high1.88 (0.75) 1.45 (0.58) 1.06 (0.49) overall1.88 (0.72)1.28 (0.64)1.14 (0.66) Tendency toward more genre diversity for openness_high Tendency toward more theme diversity for openness_low We need to repeat this study! UMAP’2013. Rome, Italy 17

18 Results (again)  No statistically significant effect of story  Effect of order in sequence  But a tendency toward a difference in application of thematic and categorical diversity UMAP’2013. Rome, Italy 18

19 Limitations  Domain  Is this what people need or what people do?  Predicting openness to experience UMAP’2013. Rome, Italy 19

20 Possible implications for recommender systems design  Start narrow go broader  Shift focus toward more thematic diversity (within cluster) for people who are low on openness to experience.  Diversity across genres (across clusters) is still relevant for the majority of users  Worth looking into predicting openness to experience UMAP’2013. Rome, Italy 20

21 Wrap-up  We studied the effect of openness to experience on rec diversity  People like to expand each other’s horizons  They start narrow and go wide  Do not really consider personality  But tends toward more thematic diversity for low OE vs more categorical diversity for low OE. UMAP’2013. Rome, Italy 21

22 Questions?   This research has been funded by the Engineering and Physical Sciences Research Council (EPSRC, UK), grant ref. EP/J012084/1 UMAP’2013. Rome, Italy 22


Download ppt "Adapting Recommendation Diversity to Openness to Experience: A Study of Human Behaviour Nava Tintarev, Matt Dennis and Judith Masthoff University of Aberdeen."

Similar presentations


Ads by Google