1 / 14

מערכות המלצה / Collaborative Filtering

מערכות המלצה / Collaborative Filtering. ד"ר אבי רוזנפלד. למצוא את הדמיון בין דברים. Supervised או Unsupervised. עדיין Unsupervised לא ברור מה ה " Objective truth " (אמת אובייקטיבי) מאוד קשה להעריך כמה טוב המערכת אחוזי קניות ( Click-through ) זמן חישוב Correlation

brina
Download Presentation

מערכות המלצה / Collaborative Filtering

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. מערכות המלצה /Collaborative Filtering ד"ר אבי רוזנפלד

  2. למצוא את הדמיון בין דברים...

  3. Supervised או Unsupervised • עדיין Unsupervised • לא ברור מה ה "Objective truth" (אמת אובייקטיבי) • מאוד קשה להעריך כמה טוב המערכת • אחוזי קניות (Click-through) • זמן חישוב • Correlation • SME (Square Mean Error)

  4. שתי שיטות כלליות • Content-based recommendations: • The user will be recommended items similar to the ones the user preferred in the past (Item based) • Collaborative filtering (or collaborative recommendations): • The user will be recommended items that people with similar tastes and preferences liked in the past (user based) • Hybrids: Combine collaborative and content-based methods. CS583, Bing Liu, UIC

  5. הרעיון הכללי • יש תלות בין ההתנהגות של אנשים: • יש אנשים שאוהבים לקנות גאדג'טים ואני יודע שאתה קנית גאדג'ט, אז מן הסתם אתה גם יקנה עוד גאדג'טים • Users Based Collaborative Filtering • יש תלות בין הדברים שבנאדם קונה • Item Based Collaborative Filtering • מי שקנה נייד מן הסתם יקנה תיק לנייד • פחות "קלאסי" אבל עדיין מאוד בשימוש • Amazon.com • http://www.cs.umd.edu/~samir/498/Amazon-Recommendations.pdf

  6. User Based Collaborative Filtering • Nearest-Neighbor CF algorithm • kNN • Cosine distance • For N-dimensional vector of items, measure two customers A and B

  7. בניית מטריצת מאפיינים

  8. Neighborhood formation phase • Let the record (or profile) of the target user be u(represented as a vector),and the record of another user be v (v T). • The similarity between the target user, u, and a neighbor, v, can be calculated using the Pearson’s correlation coefficient: CS583, Bing Liu, UIC

  9. Recommendation Phase • Use the following formula to compute the rating prediction of item i for target user u where V is the set of k similar users, rv,i is the rating of user v given to item i, CS583, Bing Liu, UIC

  10. Issue with the user-based kNN CF • The problem with the user-based formulation of collaborative filtering is the lack of scalability: • it requires the real-time comparison of the target user to all user records in order to generate predictions. • A variation of this approach that remedies this problem is called item-based CF. CS583, Bing Liu, UIC

  11. שיטות אחרות -- Clustering • Work by identifying groups of consumers who appear to have similar preferences • Performance can be good with smaller size of group • May hurt accuracy while dividing the population into clusters

  12. Item-to-Item CF Algorithm

  13. Association rule-based CF • Association rules obviously can be used for recommendation. • Each transaction for association rule mining is the set of items bought by a particular user. • We can find item association rules, e.g., buy_X, buy_Y -> buy_Z • Rank items based on measures such as confidence, etc. CS583, Bing Liu, UIC

  14. יש עוד הרבה שיטות! • תחרות הNETFLIX • http://en.wikipedia.org/wiki/Netflix_Prize • $1,000,000! • מאמר מאת יהודה קורן לגבי הניצחון (המשותף) שלו • http://dl.acm.org/citation.cfm?id=1345465 • ויש כמה שיטות היברידיות • אחד מהם הוא שלי! • Tammar Shrot, Avi Rosenfeld, Jennifer Golbeck, Sarit Kraus: CRISP: an interruption management algorithm based on collaborative filtering. CHI 2014: 3035-3044

More Related