Spam Progress And Notes

• Tag blur computation by known-good urls
  • To do this, must:
    1. Identify known good urls from DMOZ [done]
    2. Compute similarities based on these URLs' tags [in progress]
    3. Use these sims to compute tag blur for Givalink posts [code written but not run]
• Tag spam computation by known-spam urls
  • To do this, must:
    1. Compute Pr(t) for known-spam URLs in Bibsonomy [done]
    2. Compute tag_spam for Givalink posts [in progress]
• Both of the above must be inserted into Givalink database
  • To do this, must
    1. Get migrations working [Ben]
    2. Create another migration to add tables / columns where appropriate [started]
    3. Modify above scripts to output to db [not started]
• Some manual labeling of spam in Givealink
  • Identify likely spam candidates by previous two measures
  • Create page in administrator controller which allows manual spam classification