Monthly Archives: February 2016

Phil 2.4.16

7:00 – 4:00 VTX

  • The way to handle multidimensional (human) ranking of documents (i.e. web pages) is to take the dimensions and and webpages and put them on a matrix? Each page has a greater or lesser score on that dimension. Then apply page rank. Tweak weights until pages order the way we think they should
  • Does “authority” mean quality? predicting expert quality ratings of Web documents
  • LandScan (Oak Ridge Labs)
  • Uppsala Conflict Data Program Geo-referenced Event Dataset
  • Nils Weidmann Dataverse (University of Konstanz)
  • Continuing On the Accuracy of Media-based Conflict Event Data. Done. Wow. And look at all the databases ^^^ !
  • Microsoft bot API
  • Back to GoogleHacking
    • Added ‘CredEngine1’ as BASELINE search engine
    • Looks like we blew through our limits. Using my key. Verified that the BASELINE search runs. That does mean that the current 4 queries factor out to 24 searches (6 search engines * 4 queries)
    • Building search persistent object
    • Building result item object. Actually, building a JasonLoadable base class since this trick is going to be used for the query items and info object
    • Need a result info object that stores the meta information.
    • Just stumbled across a GCS twitter search. Neat.
    • Hitting the CSE and getting results. Tomorrow I’ll finish of the classes that will persist the search results. I’ve got a buffered search result to use instead of hitting google. Although it will still need to pull down the document referenced in the result. I wonder how Jsoup handles pdf and Word documents?

Phil 2.3.16

7:00 – 3:00 VTX

  • Just discovered Publius –  a Web publishing system that is highly resistant to censorship and provides publishers with a high degree of anonymity. No longer active, but produced a paper.
  • Continuing On the Accuracy of Media-based Conflict Event Data. Currently starting Matching Media-based Conflict Reports with Military Records
  • Back to Googlehacking
    • Since I’ve got the provider JSON, setting up objects that I can use for more in-depth parsing. Thinking that this could be an example of ‘code’ in the dictionary. A work can be an object that knows how to look through a section of text to see if it can find itself.
    • I think running several dictionaries over a document could be interesting. For example, using a medical and a legal dictionary on a document would let the system infer malpractice as opposed to a document on foreign aid.
    • Generating the right queries and they work in the browser:
      "Ram Singh"
      	ALL_GOV(board actions):
      	ALL_US(board actions):
      	ALL_ORG(board actions):
      	RESTRICTED_COM(board actions):
      	ALL_EDU(board actions):
  • So the next thing is to start running these queries and looking at the results to see if there are patterns. And I would be further along, but IntelliJ choked when I tried to add JPA. After flailing for a while I just gave up, created a new project, copied all the lib src and persistence directories over, updated the structure, and it all works. Grumble grumble.

Phil 2/2/16

7:00 –

Phil 2.1.16

9:00 – 4:00VTX