Discovering Temporal Hidden Contexts in Web Sessions for User Trail Prediction
•Download as PPTX, PDF•
0 likes•392 views
This document proposes discovering temporal hidden contexts in web sessions to improve user trail prediction. It formulates temporal context discovery as an optimization problem solved using hierarchical clustering. The method is demonstrated on a real-world dataset of university website sessions, discovering seven contextual categories that improve next action prediction compared to a context-unaware model. Future work includes testing on additional datasets and considering multidimensional contextual features.
Recommended
Recommended
More Related Content
Similar to Discovering Temporal Hidden Contexts in Web Sessions for User Trail Prediction
Similar to Discovering Temporal Hidden Contexts in Web Sessions for User Trail Prediction (20)
More from Julia Kiseleva
More from Julia Kiseleva (12)
Recently uploaded
Recently uploaded (20)
Discovering Temporal Hidden Contexts in Web Sessions for User Trail Prediction
- 1. Discovering temporal hidden contexts in web sessions for user trail prediction Julia Kiseleva (Eindhoven University of Technology), Hoang Thanh Lam (Eindhoven University of Technology), Mykola Pechenizkiy (Eindhoven University of Technology), Toon Calders (Université Libre de Bruxelles)
- 2. User intention prediction Search Refine Search PaymentClick Product View Click ? What is next? • Does exist any contextual information? • How we can discover it? • How we can utilize it?
- 3. Contextual Partitioning • Approaches to create local models: o Horizontal partitions Users from Europe Users from South America Session 1 Search Refine Search Click on Banner Product View Payment Session 3 Product View Payment Session 3 Search Refine Search Refine Search Click on Banner Session 4 Search Refine Search Click on Banner Product View Payment Session 5 Product View Click on Banner Search
- 4. User navigation graph Search Refine Search Payment Click on Banner Product View 1.0 2/3 1/3 1/2 1/4 Drop out 3/4 1/4 1 1/4
- 6. • Approaches to create local models: o Horizontal partition o Vertical partition : • Two types of behavior: o Ready to buy – (Product View, Payment) Session 1 Search Refine Search Click on Banner Product View Payment Session 2 Product View Payment Session 3 Search Refine Search Refine Search Click on Banner Session 4 Search Refine Search Click on Banner Product View Payment Session 5 Product View Click on Banner Search Contextual Partitioning
- 7. • Approaches to create local models: o Horizontal partition o This work is about Vertical partition: • Two types of behavior: o Ready to by – (Product View, Payment) o Just browsing – (Search, Refine Search, Click on Banner) Session 1 Search Refine Search Click on Banner Product View Payment Session 2 Product View Payment Session 3 Search Refine Search Refine Search Click on Banner Session 4 Search Refine Search Click on Banner Product View Payment Session 5 Product View Click on Banner Search Contextual Partitioning
- 8. • Approaches to create local models: o Horizontal partition o Vertical partition: • Two types of behavior: o Ready to buy – (Product View, Payment) o Just browsing – (Search, Refine Search, Click on Banner) Session 1 Search Refine Search Click on Banner Product View Payment Session 2 Product View Payment Session 3 Search Refine Search Refine Search Click on Banner Session 4 Search Refine Search Click on Banner Product View Payment Session 5 Product View Click on Banner Search Contextual Partitioning
- 9. Problem Description Timeline t5t0 t3t2 t4 t1 Search Refine Search PaymentClick Product View Search Click t6 Context ``Find information” Context ``Buy product” What is next? Probably user will change intent?
- 10. Problem Description Timeline t5t0 t3t2 t4 t1 Search = A Refine Search = B Payment Click = C Product View = D Search Click t6 Context ``Find information” Context ``Buy product”
- 11. Context-Awareness Example Contextual Feature: User intent DATA M1 M2 Contextual Categories Individual Models Mapping G Mapping H Context Discovery Ready to buy Just browsing
- 12. Context-Awareness …………… … C1 C2 C3 Cn Contextual features Fs DATA Environment M1 M2 M3 Mk Contextual Categories Individual Learners Mapping G Mapping H Context Discovery Context-Awareness: (G,H)
- 13. Temporal Context- Awareness M1 M2 Mk G H Temporal Context-Awareness: (G,H,ti) …….. t2 t1 tl-1 t3 tl C2 C1 Ck G G H H a1 a2 a3 al-1 al WebSessionS Contextual features Contextual Categories C2 Individual Models
- 14. Discovery hidden contexts Web log Train: To train predictive models Validation: To find Best clusters Test: To derive final accuracy To find a “Best” clusters Calculate final accuracy To train local predictive models
- 16. a b c d abababababcdcdababcdcdcd The number of true predictions = 0 a b c d 1.0 1.0 1.0 1.0 M1 M2 M3 M4 Hierarchical clustering
- 17. a b c d ab abababababcdcdababcdcdcd The number of true predictions = 12 a b c d 1.0 1.0 1.0 1.0 M1 M2 M3 Hierarchical clustering
- 18. a b c d ab abababababcdcdababcdcdcd The number of true predictions = 20 a b c d 1.0 1.0 1.0 1.0 cd M1 M2 Hierarchical clustering
- 19. a b c d ab abababababcdcdababcdcdcd The number of true predictions = 20 a b c d 1.0 1.0 1.0 1.0 cd abcd Stop as long as no additional prediction benefit of merging M1 M2 Hierarchical clustering
- 21. Mastersportal.eu - Search Refine Search Click on Program is Search Result Click on University Click on Country
- 22. Dataset Date May 2012 #sessions 350.618 #requests 1.775.711 #sessions from Eu 159.991
- 23. Results
- 24. Explaining the results eb C d a1.0 1.0 0.9 0.1 1.0 1.0
- 25. Explaining the results eb C d a 1.0 1.0 0.9 0.1 1.0 1.0 Explanation: 1. Users’ behavior doesn’t follow Markov property 2. Ambiguous Transition Matrix
- 26. Explaining the results University on nearby universities Related Programs Quick Search 0.306 Related Programs Related Programs Click on Program
- 27. Resulted Clusters Id Summary Cluster 1 Intensive Search Basic Search, Refine Search, Empty Search Result 2 Explore information related to program Program impression in search result, Banner click, Program click , Click on university link 3 Start of browsing University Spotlight impression, Quick search 4 Explore country information File view, Click on country link 5 Explore search result Program impressions in search results, University impression on nearby universities 6 Explore program Program in landing page, Submit inquiry 7 Outlier Submit question, X-node
- 28. Site Map • Page is represented as set of possible actions o Example: Homepage is (Quick Search, University Spot light impression, Question Submit) o Calculate Jaccard similarity between Page and Cluster
- 29. Site map Pages Clusters Homepage SearchResult ProgramPage CountryPage UniversityPage Inquiry Submission Question Submission Intensive Search Explore information related to program Start Browsing Explore country information Explore search result Explore program Outlier
- 30. Conclusion • We formulate the problem of temporal context discovery as an optimization problem. • A hierarchical clustering method is proposed to determine the optimal number of hidden contexts and mine temporal contexts. • We show a real-world use case in which the contexts as we defined them do exist and are useful for prediction.
- 31. Future works • Testing our method on another datasets • Introducing a mechanism for detecting context switching within a web-session • Considering multidimensional contextual features.
- 32. Thank you! • Context identification and integration it into prediction models • Accurately predicting users’ desired actions and understanding behavioral patterns of users in various web- applications • Personalization and adaptation to diverse customer need and preferences • Accounting for the practical needs within the considered application areas.
- 33. Conclusion • We formulate the problem of temporal context discovery as an optimization problem. • A hierarchical clustering method is proposed to determine the optimal number of hidden contexts and mine temporal contexts. • We show a real-world use case in which the contexts as we defined them do exist and are useful for prediction. Questions?
- 34. Outline • Problem description • Def. context-aw => temporal CA • Motivational example • Objectives • Method: HC • Data • Results • Summary • Conclusion • Future works
- 35. • Web sessions: • Type of events: • Web session is an ordered sequence of events: • Space of contextual features: • Predictive model: • Evaluation function: • Contextual categories:
Editor's Notes
- Add logotips
- What is the next action prediction of the user?The most common – make use of Markov Model. The question what kind of context we can define and/or discover for this problem
- In order to build a local Markov models we can partition our data.There are two types of partitioning: horizontal and vertical. Let’s start from horizontal. As context we use “user geo location”.
- We can process our historical data and build a user navigation graph. Which is suitable to train a markov models.We can incorporate context to build localMmarkov models which is easy to train and parametrized.
- Reduce an entropy.Local models are easy to learnLocal models are easy parameterized
- Let’s consider vertical partion
- For simplicicty
- Same n
- On the slide
- Add formula from page
- Masterportalhomepge to make you familiar with the portal
- Screenshot
- Explane the results
- What really context we have found
- Jaccard similarity between clusters and pagesExplore Intensive search on search result page
- Statement 1 line
- Put capa project laogand goal
- Statement 1 line