Add new item similarity metrics for SAR

[simonzhaoms]

Description

This PR adds 4 new item similarity metrics for SAR:

• Mutual information: $s_{ij} = \log_2\frac{nc_{ij}}{c_{ii}c_{jj}}$
• Lexicographers mutual information: $s_{ij} = c_{ij}\log_2\frac{nc_{ij}}{c_{ii}c_{jj}}$
• Cosine similarity: $s_{ij} = \frac{c_{ij}}{\sqrt{c_{ii}c_{jj}}}$
• Inclusion index: $s_{ij} = \frac{c_{ij}}{\min (c_{ii}, c_{jj})}$

Related Issues

• Investigate more similarity metrics for SAR for single node #126

Checklist:

• I have followed the contribution guidelines and code style for this project.
• I have added tests covering my contributions.
• I have updated the documentation accordingly.
• This PR is being made to staging branch and not to main branch.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[miguelgfierro]

Something to consider @simonzhaoms is a metric called Earth Mover's Distance I have used that in the past and worked well for me. It was for comparing feature vectors and I was using a large number of distances: euclidean, KL, Jasen-Shannon, and others. And I remember that this one got really good results. Not sure if it could be applicable in the SAR scenario

[miguelgfierro]

This is really good and interesting work Simon