There are a lot of factors that go into our search algorithm, but in brief, we crawl the web, check various sources for ratings, and we also do sentiment analysis when possible.
We're just getting started, and will refine it over time, but we had to start somewhere, and get this thing out in the world to see if it resonates with people. The jury is still out.
IMO your algorithms can be a starting point, but your results would be a lot more useful if actual humans then did a second pass and added various pros and cons to the most popular queries to enhance their trustworthiness and level of detail.
The algorithm on its own is not working well... at all. It's no better, and often worse, than generic "best of" queries in any other search engine. More details in my other post here: https://news.ycombinator.com/item?id=32500543
Comments
What ratings are you using to determine which things are "best"?
There are a lot of factors that go into our search algorithm, but in brief, we crawl the web, check various sources for ratings, and we also do sentiment analysis when possible.
We're just getting started, and will refine it over time, but we had to start somewhere, and get this thing out in the world to see if it resonates with people. The jury is still out.
IMO your algorithms can be a starting point, but your results would be a lot more useful if actual humans then did a second pass and added various pros and cons to the most popular queries to enhance their trustworthiness and level of detail.
The algorithm on its own is not working well... at all. It's no better, and often worse, than generic "best of" queries in any other search engine. More details in my other post here: https://news.ycombinator.com/item?id=32500543
We were hoping that the voting system would allow for user based manual refinement. We'll see. I'm not sure the users will be that invested.