From cb67e0fd1741f9c95343931af71f4e94a2104f34 Mon Sep 17 00:00:00 2001 From: egor-bogomolov Date: Sat, 30 May 2020 10:46:05 +0300 Subject: [PATCH] Arrange sections --- README.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index 2bcb5b5..da2236a 100644 --- a/README.md +++ b/README.md @@ -39,17 +39,17 @@ with the tool shows that both options produce similar results. It is worth notin In the end, for a given project, the tool will output its nearest neigbors among these 9 million projects, sorted by their similarity. -### Video demonstration - -We recorded a short video demonstrating the features of _Sosed_: finding similar projects, filtering them by -language and GitHub stars, producing explainable output. The video is [available on Youtube](https://www.youtube.com/watch?v=LYLkztCGRt8). - ### Explaining the output In order for *Sosed* not to act as a black-box, we manually labeled all 256 clusters with short descriptions of their topics. If you pass `--explain` flag to *Sosed*, it will identify clusters that contributed the most to the similarity and print their descriptions. +### Video demonstration + +We recorded a short video demonstrating the features of _Sosed_: finding similar projects, filtering them by +language and GitHub stars, producing explainable output. The video is [available on Youtube](https://www.youtube.com/watch?v=LYLkztCGRt8). + ## Getting started To run the tool, clone this project: