It's been exactly one year since our first launch, on 31st May 2011. Back then we had a different name "KontentLinks" and we were just at the beginning of an exciting journey. One year later, we're relaunching again, with a new mission, more focused on what we think is the biggest problem in the research community: spreading our research results.
Although nowadays there are a lot of platforms that allow people to share content, the research community has been very slow adopting Web 2.0 tools. Mendeley, Academia.edu and ResearchGate are the leading platforms in the field and they're doing a very good job. However, we think there's one thing that's constantly missing: the focus on the connections between the research works. Citations are the base for evaluating research results and although this is starting to change (i.e. altmetrics) there's no equivalent in the online space. We believe that researchers should be able to easily connect two pieces of research, creating a web of interconnected research results that can be explored by everybody, and ultimately be used for spreading new research results. And that's how Kleenk was born, a kleenk being simply a semantic connection between two scientific works which can be enriched with more details and evaluated and commented by the community.
The new Kleenk is exclusively focused on making it simple to create these connections. Key features include:
- a simple way to create kleenks
- a simple way of visualising kleenks, through automatically generated kleenk maps
- public profile page with kleenks create by you or to your work
- a simple kleenks feed and the possibility to follow researchers that you're interested in
- integration with the Mendeley platform, for easy access to your papers
Obviously creating all the connections for existing papers requires a lot of work and time from the researchers part. That's why we went one step further and tried to imagine ways of bootstrapping the kleenks database in a way that it is simple and useful for the community. Today we're also launching an experimental search feature based on the Springer API. It works in a very simple way:
- you provide your search terms
- we use them to query the Springer database, both the general one and the open access one, and we gather as much meta-data as possible, including abstracts and full texts where available
- we then analyze all the results, and identify the most important connections by:
- first analyzing the trends (the most important topics, keywords, etc) for your search;
- based on the identified importance of different types of meta-data we compute how likely two papers are to be related (i.e. how many and what keywords they have in common);
- then we also take into account some heuristic similarity algorithms which we run on titles, abstracts or full text where available;
- after all pairs of documents have been analyzed and scored, we choose the most important ones and create the connections;
- finally, we also add any existing connections created by other users; this is especially important as it allows new research results to show up next to classical papers in a research topic.
The results are presented in the form of a kleenk map which helps researchers understand faster how the results connect to each other, what's important, and finally get faster to the papers they're interested in.
Obviously this is just the beginning and there's room for lots of improvements. For example now we only identify two types of connections: 'related to' and 'strongly related to'. We're currently experimenting, for the open access papers, with using the context in which a citation is made to identify the verb (type) that best describes the relation between the two papers.
We'll be providing more details about the new features over the comming weeks so follow our blog if you want to stay updated. And if you like what we're doing please drop us a comment, we really appreciate it.
Enjoy the new Kleenk,
The Kleenk Team
