![]() ![]() adding comments or highlighting passages they consider important. In addition, Docear allows its users to make annotations in PDF files, i.e. ![]() That means Docear knows what a researcher is currently looking for, which articles a researcher already knows, which ones he is currently reading and on which new papers he is currently working. With Docear, users search for literature, they organize their literature, and they draft their own literature. Docear has access to quite diverse, and quite a lot, information about its users. For instance, on Last.fm a typical user listens to a few dozens of songs – on a single day. In contrast, recommender systems in other domains often have access to much more information about their users. For instance, if a recommender system analyses the papers a researcher has published, there are maybe a few dozens of papers to analyze. All these recommender systems suffer from one problem: They have rather limited information about their users. Other research paper recommender systems include CiteULike’s recommender system, Claper, and SCuBA. Among others, they used citations instead of words to find similar scientific articles. For the academic search engine CiteSeer there have been several different recommender systems proposed – some from the CiteSeer developers and some from third parties. Then, items in the collection of the similar user are recommended to the other user. ![]() In CF, similar users are determined by comparing how they rated items (the more often two users rated items alike, the more similar they are assumed to be). In CBF, the words of a user’s documents are taken to build a user model, and documents that contain the same words as the user model are recommended. TechLens used the two most popular recommendation approaches – content based filtering (CBF) and collaborative filtering (CF). One of the earliest research paper recommender systems was TechLens. In this paper we present the research paper recommender system which we developed for Docear. More information on Docear can be found in. the “entities” with the words) link to articles in which the information was originally found. The mind map we created outlines the skeleton of this paper and the nodes (i.e. Figure 1 shows an example of a mind map we created as draft for this paper. That means users organize their data in a tree-like data structure and not in a table or with social tags. Docear has the unique feature of utilizing mind maps for information management. Our open source tool Docear (supports researchers with literature management by bundling several applications that help in searching, organizing, and creating academic literature. In addition, full- texts are often not freely accessible and need to be paid unless ones university or library has a subscription for the publisher. Especially the search for relevant literature is challenging due to the millions of articles and books being published every year and the fact that most search services such as the ACM Digital Library focus only on publications of selected publishers (e.g. searching, organizing and creating literature, is important for researchers and students. systems, user model, mind map, mind mapping, research paper recommender system, content based filtering Literature management, i.e.
0 Comments
Leave a Reply. |