Is it time for a career switch


Published on

Published in: Technology, Education
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Is it time for a career switch

  1. 1. Is It Time For a Career Switch?
  2. 2. ABSTRACT  Tenure  what to recommend to a user  when to make appropriate recommendations and its impact on the item selection in the context of a job recommender system  The proportional hazards model, hierarchical Bayesian framework  estimates the likelihood of a user’s decision to make a job transition at a certain time, which is denoted as the tenure-based decision probability
  3. 3. CONTRIBUTION  Analyze the problem of finding the right time to make recommendations in the job domain.  Propose using the proportional hazards model to tackle the problem and extend it with a hierarchical Bayesian framework.  Evaluate the model with a real-world job application data from LinkedIn
  4. 4. RELATED WORK  Major recommendation approaches: content-based filtering and collaborative filtering  Timeliness  E.g. software engineer – senior software engineer
  5. 5. Method  HIERARCHICAL PROPORTIONAL HAZARDS MODEL   Review of Proportional Hazards Model  Model Extension with Bayesian Framework   Problem Definition Parameter Estimation TENURE-BASED DECISION PROBABILITY
  8. 8. HIERARCHICAL PROPORTIONAL HAZARDS MODEL Goal  Predict the probability that a user makes a decision of item jb at current time tb ,given that she made the last decision of ja at time ta and she did not make the transition decision up to time tb .
  9. 9. HIERARCHICAL PROPORTIONAL HAZARDS MODEL  Review of Proportional Hazards Model  Survival function :determines the failure of an event  Failure: as a user making a decision to transit to a new job
  10. 10. HIERARCHICAL PROPORTIONAL HAZARDS MODEL  Two common approaches to incorporate covariates x in the hazards model:  Cox proportional hazards model :  Accelerated life model :
  11. 11. Weibull distribution is used for p ( y ) θ :
  12. 12. Model Extension with Bayesian Framework Data sparsity
  13. 13. Parameter Estimation
  15. 15. TENURE-BASED DECISION PROBABILITY  Push-based Scenario  Pull-based Scenario
  16. 16. EVALUATION OF HAZARDS MODEL  tenure-based decision probability  predicted decision time  covariates
  17. 17. EVALUATION OF HAZARDS MODEL  Evaluation Metrics  Perplexity/Likelihood  Estimated Decision Time
  18. 18. EVALUATION OF HAZARDS MODEL  Models to Compare  H-One :the hazards model that fits a single set of parameters with no covariates ;m = {∗ → ∗}  H-Source :the hazards model that fits multiple sets of parameters with no covariates to the tenure data ;m = { a → ∗}  H-SourceDest :m = { a → b }  H-SourceDestCov : incorporates covariates into the hazards model in HSourceDest .
  19. 19. EVALUATION OF HAZARDS MODEL  covariates  1) about the user u : the user’s gender, age, number of connections, number of jobs that the user has changed, average months that the user changes a job;  2)about the item ja or jb : discretized company size, the company age  3)about the relationship between ja and jb : the ratio of the company size, the ratio of the company age; whether j a and j b are in the same function, whether they are in the same industry;  4)about the user’s aspiration of category b : number of job applications from user u in category b in the last week, last month, last two months, and last three months.
  20. 20. EVALUATION OF RECOMMENDATION MODEL  In the Push-Based Scenario
  21. 21. In the Pull-Based Scenario • BasicModel • Basic+TranProb • Basic+TranProb+Tenure • Basic+TranProb+TenureProb
  22. 22. CONCLUSION  Q: When is the right time to make a job recommendation and how do we use this inference to improve the utility of a job recommender system?  the hierarchical proportional hazards model  real-world job application data : Linkedin
  23. 23. Is This App Safe for Children? A Comparison Study of Maturity Ratings on Android and iOS Applications
  24. 24. ABSTRACT  we develop mechanisms to verify the maturity ratings of mobile apps and investigate possible reasons behind the incorrect ratings.
  25. 25. INTRODUCTION  Android maturity rating policy   “Everyone,” “Low Maturity,” “Medium Maturity,” and “High Maturity,” iOS’s policy  “4+,” “9+,” “12+,” and “17+.”  iOS rates each app submitted according to its own policies  Android apps are purely a result of app developers’ self-report.
  26. 26. INTRODUCTION  Android rating policy is unclear, and it is difficult for developers to understand the difference between the four maturity-rating levels  Contribution:  We develop a text mining algorithm to automatically predict apps’ actual maturity ratings from app descriptions and user reviews.  By comparing Android ratings with iOS ratings, we illustrate the percentage of Android apps with incorrect maturity ratings and examine the types of apps which tend to be misclassified.  We conduct some preliminary analyses to explore the factors that may lead to untruthful maturity ratings in Android apps.
  27. 27. RESEARCH QUESTIONS  Does iOS rating strictly reflect its policy?  Are app ratings reflected in app descriptions and user reviews? If so, can we build an effective text mining approach to predict the true rating of an app?  Do Android developers provide accurate maturity ratings for their own apps? For apps published in both markets, are Android ratings consistent with iOS ratings?  What are the factors that could lead to untruthful maturity ratings in Android apps in comparison to iOS apps?
  28. 28. METHODOLOGY  iOS Maturity Rating Policy vs. Implementation
  29. 29.  iOS actually downgrades its official maturity policy during implementation.
  30. 30. Android Apps’ Maturity Ratings
  31. 31. discrepancies • Android does not consider horror content ( C ) as mature content, while iOS does include • Android considers graphic violence ( B3 ) as mature content while iOS directly rejects apps with graphic violence. • Android integrates privacy protection in its maturity rating policy by including the social feature ( I ) and location collection ( J ). However, no corresponding privacyrelated consideration exists in the maturity rating scheme by iOS. • Frequent/intense cartoon violence and fantasy violence ( A2 ) is rated as “Medium Maturity” (i.e., level 3) in Android but as “9+” (i.e., level 2) in iOS. • Frequent/intense simulated gambling ( H2 ) is rated as “High Maturity” (i.e., level 4) in Android but is rates as “12+” (i.e., level 3) in iOS.
  32. 32. we can now use iOS actual maturity rating as a baseline to examine the reliability of Android apps’ maturity ratings.
  33. 33. Comparing Apps on iOS and Android  For each Android app, we choose up to 150 search results from the iOS App Store. For those showing similar app names, we conducted analysis to determine the closest fit.  their descriptions and developers’ company names  apps’ icons and screenshots
  34. 34. ALM—Automatic Label of Maturity Ratings for Mobile Apps  “Android-only” apps  ALM is a semi-supervised learning algorithm, and it processes apps’ descriptions and user reviews to determine maturity ratings. 1. Building seed-lexicons for objectionable content detection 2. Assigning initial weights to seed-terms 3. Classification 4. Expanding seed-lexicons and adjusting weights
  35. 35. 1.Building seed-lexicons for objectionable content detection  Apps are organized based on their rating scheme together with their corresponding token, such as A1.txt , A2.txt , B1.txt , and H2.txt .  Human experts read grouped app descriptions and select seed lexicons to detect objectionable content.  grouped into three bigger lexicons denoted as Ti, i∈ 9,12,17 for classifying the maturity rating: 9+, 12+, and 17+
  36. 36. 2.Assigning initial weights to seedterms Pi, Ni For each seed-term t, denote its frequency in Pi and Ni as tp and tn
  37. 37. 2.Assigning initial weights to seedterms
  38. 38. 3.Classification  For each app 4 , all terms in its description are selected and categorized as a set A=tk  maturity rating ma
  39. 39. 4.Expanding seed-lexicons and adjusting weights
  40. 40. EXPERIMENT A total of 1,464 apps were found on iOS App Store and the rest 3,595 apps were classified as Android-only apps.
  41. 41. Experiment 1: Predicting Apps’ Maturity Ratings by the ALM algorithm
  42. 42. Experiment 2: Overrated and Underrated Android Applications • overrated • underrated
  43. 43. Overrated Android Applications  possible reasons  Intelligence  Simulated Gambling  Violence  Mature and Suggestive Themes
  44. 44. Underrated Android Applications
  45. 45. Experiment 3: Exploring Factors Contributing to Incorrect Ratings  apps’ attributes : popularity, price, and dangerous level of the required permissions .  Developers’ attributes : general privacy awareness, trustworthiness, actual privacy awareness, and child safety awareness .
  46. 46. Conclusion  examine the maturity rating policies on both Android and iOS platforms  possible reasons behind the incorrect ratings  ALM algorithm