Interpretability of the Similarity Model and Scores

@fxiafxia and I were never quite clear on how exactly the similarity model for this worked, or how to interpret the scores. We thought this could be addressed a little more in the readme

Start with the README and locate the existing explanations of the similarity model and its scores. Compare the documentation with the project's current behavior, then expand the README so a newcomer can understand how the model works and interpret its scores.

索引モデルが issue の本文から書いたものです。