Addressing Cold-Start in App Recommendation: Latent User Models Constructed from Twitter Followers

By Jovian Lin, Kazunari Sugiyama, Min-Yen Kan, and Tat-Seng Chua.

Presented at the 36th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR'13), Dublin, Ireland, July 28--August 1, 2013.

Source available at: http://jovianlin.com

#SIGIR #SIGIR13 #SIGIR2013 #NUS #WING-NUS

Transcript

1. Addressing Cold-Start in App Recommendations:
   Jovian Lin / Kazunari Sugiyama / Min-Yen Kan / Tat-Seng Chua
   National University of Singapore
   Latent User Models Constructed from Twitter Followers
   

   View Slide

2. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
   

   View Slide

3. Many, many apps.
   Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
   No. of
   apps
   Time
   1/32
   

   View Slide

4. View Slide

5. INFORMATION
   OVERLOAD
   !
   

   View Slide

6. Content-based Filtering
   Recommender Systems
   Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
   Collaborative Filtering
   2/32
   

   View Slide

7. Content-based Filtering
   Recommender Systems
   Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
   Collaborative Filtering
   4 1 4
   5 4 ?
   1 ? 5
   ? 5 ?
   1 ? 4
   items
   2
   1 3
   1
   2
   3
   4
   5
   users
   ? ?
   ? ?
   ? ?
   ? ?
   ? ?
   4 5
   new items
   3/32
   

   View Slide

8. Content-based Filtering
   Recommender Systems
   Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
   Collaborative Filtering
   4 1 4
   5 4 ?
   1 ? 5
   ? 5 ?
   1 ? 4
   items
   2
   1 3
   1
   2
   3
   4
   5
   users
   ? ?
   ? ?
   ? ?
   ? ?
   ? ?
   4 5
   new items
   COLD START PROBLEM
   Solution #1: Wait for ratings to come in.
   Solution #2: Use content-based filtering.
   3/32
   

   View Slide

9. Collaborative Filtering Content-based Filtering
   Recommender Systems
   Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
   •  Recommends items based on similar content.
   (e.g., genres, textual descriptions)
   •  Con: Lack of diversified recommendations.
   Example: a user who has downloaded a
   weather app will receive weather-related
   app-recommendations.
   4/32
   

   View Slide

10. Meanwhile…
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion 5/32
    

    View Slide

11. View Slide

12. View Slide

13. “Can we merge information mined
    from social networks to enhance
    (app) recommendations?”
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion 6/32
    

    View Slide

14. “Can we address the cold-start in
    recommender systems by using
    nascent signals in social networks?”
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion 7/32
    

    View Slide

15. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion 8/32
    

    View Slide

16. We made two observations:
    1.  Apps contain references to their Twitter accounts.
    2.  Early signals about apps can be present in social
    networks, even before ratings are received.
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    

    View Slide

17. We made two observations:
    1.  Apps contain references to their Twitter accounts.
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion 9/32
    

    View Slide

18. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion 9/32
    

    View Slide

19. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion 9/32
    

    View Slide

20. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    Follow @angrybirds on Twitter
    9/32
    

    View Slide

21. View Slide

22. We made two observations:
    1.  Apps contain references to their Twitter accounts.
    2.  Early signals about apps can be present in social
    networks, even before ratings are received.
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion 10/32
    

    View Slide

23. We made two observations:
    1.  Apps contain references to their Twitter accounts.
    2.  Early signals about apps can be present in social
    networks, even before ratings are received.
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    Evernote iOS app
    Release Date: 8 May 2012
    May 2012 Jun 2012 Jul 2012 Dec 2012
    …
    0 ratings 0 ratings
    First few ratings
    start coming in
    118,827
    ratings
    Has an account
    on Twitter
    since Feb 2008
    By May 2012, Evernote’s
    Twitter account already
    had 120,000 followers and
    1,300 tweets.
    10/32
    

    View Slide

24. We made two observations:
    1.  Apps contain references to their Twitter accounts.
    2.  Early signals about apps can be present in social
    networks, even before ratings are received.
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    We *CAN* address the cold-start in
    recommender systems by using
    nascent signals in social networks.
    11/32
    

    View Slide

25. We want to estimate the probability that
    “a target user u will like an app a.”
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    “like” app user
    p( + | a, u )
    12/32
    

    View Slide

26. We want to estimate the probability that
    “a target user u will like an app a.”
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    p( + | a, u ) p( + | t, u) p( t | a)
    ∑
    =
    “like” app user Twitter-follower
    “Pseudo-documents” & “Pseudo-words”
    Uniform distribution over the
    various Twitter-followers (t)
    following app a.
    t∈T(a)
    13/32
    

    View Slide

27. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    downloaded/consumed
    14/32
    

    View Slide

28. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    User u
    downloads
    14/32
    

    View Slide

29. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    User u
    downloads
    twitter.com/angrybirds
    twitterID_31230
    twitterID_2289
    twitterID_999
    twitterID_50401
    ……
    followers
    twitterID_2
    twitterID_3142439
    twitterID_111031
    14/32
    

    View Slide

30. View Slide

31. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    User u
    downloads
    twitter.com/angrybirds
    twitterID_31230
    twitterID_2289
    twitterID_999
    twitterID_50401
    ……
    followers
    twitterID_2
    twitterID_3142439
    twitterID_111031
    14/32
    

    View Slide

32. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    User u
    downloads
    twitterID_31230
    twitterID_2289
    twitterID_999
    twitterID_50401
    ……
    followers
    twitterID_2
    twitterID_3142439
    twitterID_111031
    14/32
    

    View Slide

33. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    User u
    downloads
    twitterID_31230
    twitterID_2289
    twitterID_999
    twitterID_50401
    ……
    followers
    twitterID_2
    twitterID_3142439
    twitterID_111031
    14/32
    

    View Slide

34. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    User u twitterID_31230
    twitterID_2289
    twitterID_999
    twitterID_50401
    ……
    twitterID_2
    twitterID_3142439
    twitterID_111031
    14/32
    

    View Slide

35. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    User u
    twitterID_31230
    twitterID_2289
    twitterID_999
    twitterID_50401
    ……
    twitterID_2
    twitterID_3142439
    twitterID_111031
    Pseudo-Document Pseudo-Words
    15/32
    

    View Slide

36. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    Pseudo-document u
    (twitterID10
    , DISLIKED)
    (twitterID12
    , DISLIKED)
    (twitterID10
    , LIKED)
    (twitterID12
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID31
    , LIKED)
    User u
    disliked
    liked
    liked
    Followed by:
    •  twitterID10
    •  twitterID12
    Followed by:
    •  twitterID10
    •  twitterID12
    •  twitterID29
    App a
    App b
    App c
    Twitter-follower ID
    Preference indicator
    Followed by:
    •  twitterID29
    •  twitterID31
    16/32
    

    View Slide

37. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    Pseudo-document u
    (twitterID10
    , DISLIKED)
    (twitterID12
    , DISLIKED)
    (twitterID10
    , LIKED)
    (twitterID12
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID31
    , LIKED)
    Twitter-follower ID
    Preference indicator
    The concept of “pseudo-documents” and “pseudo-words”
    *does not* apply exclusively to Twitter followers.
    17/32
    

    View Slide

38. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    Pseudo-document u
    (genreID10
    , DISLIKED)
    (genreID12
    , DISLIKED)
    (genreID10
    , LIKED)
    (genreID12
    , LIKED)
    (genreID29
    , LIKED)
    (genreID29
    , LIKED)
    (genreID31
    , LIKED)
    Genre ID
    Preference indicator
    The concept of “pseudo-documents” and “pseudo-words”
    *does not* apply exclusively to Twitter followers.
    17/32
    

    View Slide

39. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    Pseudo-document u
    (wordID10
    , DISLIKED)
    (wordID12
    , DISLIKED)
    (wordID10
    , LIKED)
    (wordID12
    , LIKED)
    (wordID29
    , LIKED)
    (wordID29
    , LIKED)
    (wordID31
    , LIKED)
    Word ID
    Preference indicator
    The concept of “pseudo-documents” and “pseudo-words”
    *does not* apply exclusively to Twitter followers.
    17/32
    

    View Slide

40. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    Pseudo-document u
    (twitterID10
    , DISLIKED)
    (twitterID12
    , DISLIKED)
    (twitterID10
    , LIKED)
    (twitterID12
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID31
    , LIKED)
    17/32
    

    View Slide

41. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    Pseudo-document u
    (twitterID10
    , DISLIKED)
    (twitterID12
    , DISLIKED)
    (twitterID10
    , LIKED)
    (twitterID12
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID31
    , LIKED)
    Constructing Latent Groups
    18/32
    

    View Slide

42. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    Constructing Latent Groups
    (twitterID10
    , DISLIKED)
    (twitterID12
    , DISLIKED)
    (twitterID10
    , LIKED)
    (twitterID12
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID31
    , LIKED)
    Pseudo-documents
    LDA
    Per-document topic
    distribution
    Per-topic word
    distribution
    19/32
    

    View Slide

43. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    Constructing Latent Groups
    (twitterID10
    , DISLIKED)
    (twitterID12
    , DISLIKED)
    (twitterID10
    , LIKED)
    (twitterID12
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID31
    , LIKED)
    Pseudo-documents
    LDA
    = ∑ p( +, t | z) p( z | u)
    p( + | t, u)
    Per-document topic
    distribution
    Per-topic word
    distribution
    z∈Z
    19/32
    

    View Slide

44. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    Constructing Latent Groups
    (twitterID10
    , DISLIKED)
    (twitterID12
    , DISLIKED)
    (twitterID10
    , LIKED)
    (twitterID12
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID31
    , LIKED)
    Pseudo-documents
    LDA
    = ∑ p( +, t | z) p( z | u)
    p( + | t, u)
    Per-document topic
    distribution
    Per-topic word
    distribution
    Per-topic word
    distribution
    Per-document topic
    distribution
    z∈Z
    19/32
    

    View Slide

45. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    Constructing Latent Groups
    (twitterID10
    , DISLIKED)
    (twitterID12
    , DISLIKED)
    (twitterID10
    , LIKED)
    (twitterID12
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID29
    , LIKED)
    (twitterID31
    , LIKED)
    Pseudo-documents
    LDA
    = ∑ p( +, t | z) p( z | u)
    p( + | t, u)
    Per-document topic
    distribution
    Per-topic word
    distribution
    Probability that the presence
    of Twitter-follower t indicates
    that it is “liked” by user u.
    z∈Z
    19/32
    

    View Slide

46. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    p( + | t, u) p( t | a)
    20/32
    

    View Slide

47. We want to estimate the probability that
    “a target user u will like an app a.”
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    p( + | a, u ) p( + | t, u) p( t | a)
    ∑
    =
    “like” app user
    Uniform distribution over the
    various Twitter-followers (t)
    following app a.
    Probability that the presence
    of Twitter-follower t indicates
    that it is “liked” by user u.
    Derived from Pseudo-Documents
    and Pseudo-Words.
    t∈T(a)
    21/32
    Stated
    earlier in
    Slide 13/32
    

    View Slide

48. Dataset
    •  We collected data from the Apple iTunes Store and
    Twitter during September to December 2012.
    •  Stats:
    •  1,289,668 ratings
    •  7,116 apps (with Twitter accounts)
    •  10,133 users.
    •  Restrictions:
    •  Each user must give at least 10 ratings for apps.
    •  Each Twitter ID is related to at least 5 apps.
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion 22/32
    

    View Slide

49. Simulating Cold-Start
    •  10-fold cross validation.
    •  Selected 10% of the apps to be the held out set for
    all users.
    •  Each user has the same within-fold apps.
    •  Guarantee that none of these apps are in the
    training set of any user.
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion 23/32
    

    View Slide

50. Evaluation Metric
    •  Our system outputs M apps for each user, sorted by
    their probability of liking the apps.
    •  Recall@M
    •  Zero ratings are uncertain – it is difficult to accurately
    compute precision.
    •  Since the ratings are true positives, recall is a more
    pertinent measure – it only considers the positively
    rated apps.
    Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion 24/32
    

    View Slide

51. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    3 Research Questions (RQ)
    25/32
    

    View Slide

52. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ1: How does the performance of Twitter-followers
    feature compare with other features?
    RQ2: How does our method compare with other
    techniques?
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    25/32
    

    View Slide

53. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ1: How does the performance of Twitter-followers
    feature compare with other features?
    25/32
    

    View Slide

54. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ1: How does the performance of Twitter-followers
    feature compare with other features?
    RQ2: How does our method compare with other
    techniques?
    25/32
    

    View Slide

55. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ1: How does the performance of Twitter-followers
    feature compare with other features?
    RQ2: How does our method compare with other
    techniques?
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    25/32
    

    View Slide

56. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ1: How does the performance of Twitter-followers
    feature compare with other features?
    0
    0.1
    0.2
    0.3
    0.4
    0.5
    0.6
    0.7
    20 40 60 80 100 120 140 160 180 200
    Recall
    Number of recommended apps (M)
    Pseudo-Docs (W) Pseudo-Docs (D) Pseudo-Docs (G)
    Pseudo-Docs (T) Pseudo-Docs (All)
    •  Words (W)
    •  Developers (D)
    •  Genres (G)
    •  Twitter-followers (T)
    •  All features (All)
    All =
    T =
    G =
    D =
    W =
    All features
    Twitter-followers
    Genres
    Developers
    Words
    26/32
    

    View Slide

57. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ1: How does the performance of Twitter-followers
    feature compare with other features?
    Feature R@100
    All features (TGDW) * 0.513
    All, excluding Twitter-followers (GDW) 0.452
    All, excluding Genres (TDW) 0.491
    All, excluding Developers (TGW) 0.498
    All, excluding Words (TGD) 0.507
    Twitter-followers (T) 0.478
    Genres (G) 0.435
    Developers (D) 0.395
    Words (W) 0.373
    27/32
    

    View Slide

58. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ1: How does the performance of Twitter-followers
    feature compare with other features?
    Feature R@100
    All features (TGDW) * 0.513
    All, excluding Twitter-followers (GDW) 0.452
    All, excluding Genres (TDW) 0.491
    All, excluding Developers (TGW) 0.498
    All, excluding Words (TGD) 0.507
    Twitter-followers (T) 0.478
    Genres (G) 0.435
    Developers (D) 0.395
    Words (W) 0.373
    27/32
    

    View Slide

59. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ1: How does the performance of Twitter-followers
    feature compare with other features?
    Feature R@100
    All features (TGDW) * 0.513
    All, excluding Twitter-followers (GDW) 0.452
    All, excluding Genres (TDW) 0.491
    All, excluding Developers (TGW) 0.498
    All, excluding Words (TGD) 0.507
    Twitter-followers (T) 0.478
    Genres (G) 0.435
    Developers (D) 0.395
    Words (W) 0.373
    27/32
    

    View Slide

60. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ1: How does the performance of Twitter-followers
    feature compare with other features?
    Feature R@100
    All features (TGDW) * 0.513
    All, excluding Twitter-followers (GDW) 0.452
    All, excluding Genres (TDW) 0.491
    All, excluding Developers (TGW) 0.498
    All, excluding Words (TGD) 0.507
    Twitter-followers (T) 0.478
    Genres (G) 0.435
    Developers (D) 0.395
    Words (W) 0.373
    27/32
    

    View Slide

61. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ1: How does the performance of Twitter-followers
    feature compare with other features?
    Feature R@100
    All features (TGDW) * 0.513
    All, excluding Twitter-followers (GDW) 0.452
    All, excluding Genres (TDW) 0.491
    All, excluding Developers (TGW) 0.498
    All, excluding Words (TGD) 0.507
    Twitter-followers (T) 0.478
    Genres (G) 0.435
    Developers (D) 0.395
    Words (W) 0.373
    27/32
    

    View Slide

62. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ1: How does the performance of Twitter-followers
    feature compare with other features?
    28/32
    

    View Slide

63. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ2: How does our method compare with other
    techniques?
    0
    0.1
    0.2
    0.3
    0.4
    0.5
    0.6
    0.7
    20 40 60 80 100 120 140 160 180 200
    Recall
    Number of recommended apps (M)
    Full Dataset
    VSM (Words) VSM (Twitter) LDA
    CTR Pseudo-Docs (Twitter)* Pseudo-Docs (All)**
    •  VSM (Words)
    •  VSM (Twitter)
    •  LDA
    •  Collaborative Topic Regression
    •  Pseudo-Docs (Twitter)
    •  Pseudo-Docs (All)
    Pseudo-Docs (All)
    Pseudo-Docs (Twitter)
    CTR
    LDA
    VSM (Twitter)
    VSM (Words)
    29/32
    

    View Slide

64. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ2: How does our method compare with other
    techniques?
    0
    0.1
    0.2
    0.3
    0.4
    0.5
    20 40 60 80 100 120 140 160 180 200
    Recall
    Number of recommended apps (M)
    Limited to 15 apps per user (Sparse Dataset)
    VSM (Words) VSM (Twitter) LDA
    CTR Pseudo-Docs (Twitter)* Pseudo-Docs (All)*
    Pseudo-Docs (All)
    Pseudo-Docs (Twitter)
    CTR
    LDA
    VSM (Twitter)
    VSM (Words)
    30/32
    

    View Slide

65. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    

    View Slide

66. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    

    View Slide

67. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    

    View Slide

68. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    

    View Slide

69. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    

    View Slide

70. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    “The Cat in the Hat” (Book)
    “Christmas Cutie” (Book)
    “Books” / “Education” / “Games”
    31/32
    

    View Slide

71. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    “Friendly Shapes” (Education)
    “There’s No Place Like Space” (Education)
    “Books” / “Education” / “Games”
    31/32
    

    View Slide

72. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    “Pasta Crazy Chef” (Games)
    “Gingerbread Dress” (Games)
    “Books” / “Education” / “Games”
    31/32
    

    View Slide

73. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    Top 5 Twitter Profiles in Latent Group 1.
    “Pasta Crazy Chef” (Games)
    “Gingerbread Dress” (Games)
    “Books” / “Education” / “Games”
    31/32
    

    View Slide

74. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    Nosy Crow Apps
    Nosy Crow creates children’s books and
    apps. You may know our 3-D Fairytale
    apps, The Three Little Pigs & Cinderella.
    “Pasta Crazy Chef” (Games)
    “Gingerbread Dress” (Games)
    “Books” / “Education” / “Games”
    31/32
    

    View Slide

75. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    The iMums
    Four mums dedicated to reviewing apps
    and technology products for children to
    help educate their parents about the
    variety available.
    “Pasta Crazy Chef” (Games)
    “Gingerbread Dress” (Games)
    “Books” / “Education” / “Games”
    31/32
    

    View Slide

76. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    Mums with Apps
    Supporting family-friendly developers
    seeking to promote quality apps for kids
    and families.
    “Pasta Crazy Chef” (Games)
    “Gingerbread Dress” (Games)
    “Books” / “Education” / “Games”
    31/32
    

    View Slide

77. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    Charly James
    Div. Mom of 2 with varying SN & medical
    d/x. dandelion moms.
    “Pasta Crazy Chef” (Games)
    “Gingerbread Dress” (Games)
    “Books” / “Education” / “Games”
    31/32
    

    View Slide

78. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    Next is Great
    We create and develop brain teasing
    educational iOS apps for kids and
    teenagers. Check out Pirate Trio
    Academy and Geek Kids.
    “Pasta Crazy Chef” (Games)
    “Gingerbread Dress” (Games)
    “Books” / “Education” / “Games”
    31/32
    

    View Slide

79. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    

    View Slide

80. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    “BeatStudio”, “AmpKit+”, “GuitarStudio”,
    “Everyday Looper”, “Mixr DJ”, etc.
    “Music”
    

    View Slide

81. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    “BeatStudio”, “AmpKit+”, “GuitarStudio”,
    “Everyday Looper”, “Mixr DJ”, etc.
    “Music”
    Top 5 Twitter Profiles in Latent Group 2.
    

    View Slide

82. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    “BeatStudio”, “AmpKit+”, “GuitarStudio”,
    “Everyday Looper”, “Mixr DJ”, etc.
    “Music”
    Derek Jones
    Indie music publishing label, studio &
    brand. Blues&Rock, Progressive&Funk,
    Jazz&Fusion, Alternative&Christian,
    Classical, Education & a lot in-between
    too!
    

    View Slide

83. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    “BeatStudio”, “AmpKit+”, “GuitarStudio”,
    “Everyday Looper”, “Mixr DJ”, etc.
    “Music”
    Chip Boaz
    I’m a musician based in the San
    Francisco Bay Area with an interest in
    using my iPad, iPhone, & iPod to make
    music. Follow my iOS adventures.
    

    View Slide

84. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    “BeatStudio”, “AmpKit+”, “GuitarStudio”,
    “Everyday Looper”, “Mixr DJ”, etc.
    “Music”
    Dave Gibson
    Creator of MicroTrack dB, a music
    making app for iOS and Samsung bada.
    Musician, writer, audio engineer and
    synth nerd.
    

    View Slide

85. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    “Paper Monsters”, “Stickman Cliff Diving”,
    “Lili”, “Snoopy’s Street Fair”, “Gizmonauts”,
    etc.
    “Games”, “Photo & Video”
    

    View Slide

86. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    “Paper Monsters”, “Stickman Cliff Diving”,
    “Lili”, “Snoopy’s Street Fair”, “Gizmonauts”,
    etc.
    “Games”, “Photo & Video”
    “Video games warrior, lover of life,
    eternal student of the universe…”
    “I’m a Multimedia developer working at
    Kent State Uni! I also do art services for
    the game industry…”
    “Agalog Games is an independent iOS
    game studio…”
    “Hi! Samadhi Games LLC is an Indie
    Developer of iOS, Android, etc”
    

    View Slide

87. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    RQ3: Do the latent groups make any sense? What can
    we learn from them?
    31/32
    

    View Slide

88. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    •  Alleviate the cold-start in app-recommendation by using
    Twitter profiles of apps + Twitter followers.
    •  By using the feature of Twitter-followers to generate latent
    groups, our method works well – especially in a domain
    with unreliable textual features.
    •  Allows us to map users from the App Store to users in
    Twitter.
    •  Future work:
    •  Explore second-degree relationships on Twitter.
    •  Explore the use of our approach in other domains, such as
    music-recommendation.
    32/32
    

    View Slide

89. Thank you
    

    View Slide

90. Introduction ›❯ Our Approach ›❯ Experiments ›❯ Results ›❯ Conclusion
    •  Ratings lag behind Twitter followers.
    •  Because it takes more effort to post a rating/review
    than to follow a Twitter account.
    •  Monitored a few new apps:
    •  Average # of new ratings/reviews in a week: 4.2
    •  Average # of new Twitter followers in a week: 21.4
    •  We want to recommend ASAP – even if it’s 1 day
    faster.
    #extra
    

    View Slide