1

I am doing a CV short listing project for a company. I have a concept map for whole company documents. And also i have extracted the data from CV and now I have a concept map for each CV. I want to compare each CV with the Company Concept map for getting relevance order with company. Can anyone plz tell me about concept map comparison algorithm???

  • 1
    Post some examples, and explain how you want the two documents compared. – hyleaus Nov 21 '12 at 03:13
  • This is a sample company concept map http://img441.imageshack.us/img441/5278/69585982.jpg this is a sample cv concept map http://img696.imageshack.us/img696/8052/71021599.jpg What i want is to get similarity measure?? –  Nov 21 '12 at 03:58
  • Ways things can be similar: number of words, usages of verb tense, frequency of noun phrases, colors, file size, whether there are pictures of monkeys... – hyleaus Nov 21 '12 at 22:12

2 Answers2

0

When talking about comparison you generally need a distance/similarity measure. It will be very helpful if you can tell what you are trying to compare and whether you have some similarity measure already defined.

DotPi
  • 3,977
  • 6
  • 33
  • 53
0

This is a rather old question, but maybe my suggestions help someone else. In this specific example it is not possible to see all elements of company CM, so this suggestion may be incorrect.

The assumption is that both CMs are stored in text format, in the form of a graph or tree. My idea is simple counting (using brute force) those concepts that appear in both CMs. In CV map you can determine specific weight of concepts in individual branches - depending on company needs. E. g. overlapping of concepts in "Extra Curricular Activities" can be multiplied by value 0.25, while overlapping of concepts in "Technology" can be multiplied by value 0.75.

When counting, I would start form concepts on the 3rd level of CV map, because as I can see, concepts on the first two levels do not exist in company CM. You search through all branches of the company CM tree, and count each occurrence of concepts in both maps. The total value of each CV CM is equal to the sum of overlapping (normalized with "value" of specific CV branch). At the end just simply sort CVs by their calculated total value.

Kruno0
  • 1
  • 1