Course talk:CPSC522/Record Linkage and identity uncertainty

Hi Dandan,
It was an interesting subject, very well presented and described. Here are my suggestions:

1. A link’s format in the “Possible Solution” needs to be fixed.
2. I suggest you to apply a reference style in “References” section and refer to those last two abandoned papers in the body of page that help readers to understand what are their associations with what you explained across the page.
3. It would be great if you add your own evaluations of the two papers in a few sentences as Adnan already mentioned.
4. The page has a kind of instructing tone rather than descriptive. Personally, I would appreciate a more clear boundary between the materials adapted from each paper, although, the current structure is also expressive.

All above matters aside, it was a nice draft and I enjoyed reading it.

Good job,

Yaashaar

Hi Dandan,
Good work so far! Interesting topic and a generally well written page. I do have a few suggestions:

• As the others have already mentioned, revising your abstract would be a good idea. Right now, it doesn't really tell you what the page covers.
• You could also add your own thoughts on how successful it was and how it can be improved.

Hope this helps,
Adnan

Hi Dandan,

Good job!

Here are some suggestions:

The abstract does not cover the contents. I think it is better to expand the abstract and bring a brief abstract of the algorithms and the motivations.

I think it was better to explicitly mention the name of the papers and what they proposed and how the authors evaluate their models. What I see in the page covers lots of things, but not explicitly separate the papers,

Cheers,

Bahare

Hi Bahare,

Thank you for your suggestions.

Since we are working on a same topic, so I am trying not to overlap the part you mentioned in your page, and referenced a lot from yours. :) I would try to separate the two papers, and add some comparison details on to it.

Hi Dandan,

Great page! Learnt a lot. I liked the problem description section. Few suggestions

1. You can write a bit more in the abstract to give a comprehensive idea.
2. There are some typos.
3. A few links to other pages for some concepts and papers where the algorithms mentioned are present.
4. You can increase the size of the last image to improve readability.

Some questions I had :

1. What does x,X represent in trail(x,X) in the Location-Based Record Linkage Algorithms?
2. Similarly what exactly is supertrail in this algo?

Hi SamprityKashyap,

Thank you for your suggestions, and I would add more actual info to explain how the trail match works, the comparison with the old algorithms.