match score

How are match scores calculated?

ARRR calculates a score and a density value for each category and for the resume as a whole.

Density

Density is simply the percentage of total words in the resume which are keywords. My experience (with the example keywords file included in the download) is that good resumes come in at around 5% to 8% density.

Match Score

ARRR also calculates a match score for the resume. This is a little more complex but the in short the score is based on the number of keywords found.

Category Weighting

Each category has a weighting factor in the keywords.js file. This weighting can be used to adjust the importance of the category for the overall score.

In the example file, I care about people with Zappos experience. So I set the weighting to 3 for that category to make it very important. Conversely, there are a lot of keywords in the Tech category so to prevent it from being overly expensive I reduced the category weighting.

Diminishing scoring

The first time ARRR finds a particular keyword it increases the match score by 0.7. Every additional match is worth less and less.

This is done because if ARRR sees the keyword 'SQL' in the resume twice that is important. But if it sees the keyword twenty times the score shouldn't increase by 20!

The formula for calculating the match score of a particular keyword is: log(match-count) / log(2.5). That results in the following curve.