Skip to content

Meeting 2: [10\1\2014]

NerdPollution edited this page Oct 3, 2014 · 1 revision
  • Take aways:
    ⋅⋅⋅ Re-ranking algorithm:
    ⋅⋅⋅⋅⋅⋅ for filtering things after the stop words are removed
    ⋅⋅⋅ Add the constant:
    ⋅⋅⋅⋅⋅⋅ ???
    ⋅⋅⋅ Word constraint weak:
    ⋅⋅⋅⋅⋅⋅ Document constraints are stronger

  • Data Set:
    ⋅⋅⋅ 20 newsgroup dataset
    ⋅⋅⋅⋅⋅⋅ ground truth

  • Dynamic visualization:
    ⋅⋅⋅ static will not be efficient enough to display the level of information needed

  • Document order doesn't matter:
    ⋅⋅⋅ essentially documents are bag of words
    ⋅⋅⋅ multiple documents are needed in order for the algorithm to work
Clone this wiki locally