![]() title and journal name), and displays metadata when the mouse hovers over a PDF icon. ![]() Docear also extracts metadata from PDF files (e.g. Docear imports annotations (comments, highlighted text, and bookmarks) made in the PDFs, and clicking a PDF icon opens the linked PDF file. We created categories reflecting our research interests (“Academic Search Engines”), subcategories (“Google Scholar”), and sorted PDFs by category and subcategory. The mind-map in Figure 2 was created to manage academic PDF files and references. Some examples of mind-maps are shown in Figure 2, Figure 3, and Figure 4, which we created with Docear. In some cases, mind-mapping tools offer additional features such as the adding of images, notes, or formulas. Nodes typically contain a few terms and may link to websites or files on the users’ hard drives. To detail the child-nodes, users create child-nodes for the child-nodes, and so on. To detail the central concept, users create sub-nodes branching from the root node. When users create a new mind-map, they start with a root node in which they write the central concept that the mind-map is about (Davies 2011). Mind-maps are similar to outlines and consist of three elements: nodes, connections, and visual clues. websites in which users can upload their mind-maps for other users to browse and download 123. Around 300,000 mind-maps are publicly available in so- called “mind -map galleries, ” i.e. These tools are used by an estimated two million users who create around five millions mind-maps every year (Beel, Langer, Genzmehr, et al. Originally, mind-mapping was done with pen and paper, but since the 1980s, more than one-hundred software tools for aiding users in creating mind-maps have evolved (Gee 2012). ![]() As such, mind-maps are used for tasks like brainstorming, knowledge management, note taking, presenting, project planning, decision making, and career planning (Frey 2010). Mind-mapping is a technique used to record and organize information, and to develop new ideas (Holland et al. While others make use of tables or social tags to manage PDFs and references, Docear primarily uses mind-maps. ![]() It is comparable to reference managers such as Zotero or Endnote. Docear is open-source desktop software used for organizing references and PDFs (Beel et al. For more details on the algorithms and the evaluation, refer to Lommatzsch (2014). In contrast to an online evaluation, the near-to-online evaluation allowed us to benchmark several different approaches simultaneously, enabling us to efficiently measure the effectiveness of different approaches and parameters. Based on this definition, the expected precision grows with the number of suggestions in the predicted list. We computed precision by counting an answer as correct if the list of suggestions contained the correct article. The predicted article lists were checked against the clicks observed in the online evaluation. When a click event was received, we sent a request to one (or more) recommendation approaches and asked for predictions of what articles the user would have clicked. We analyzed the stream of click events provided by Plista. When Plista requested recommendations, Plista provided information such as a list of recommendations a user had seen before and which recommendations the user had previously clicked. Effectiveness was measured with precision in a “near to- online” evaluation. By “performing well, ” we mean that if one algorithm was the most effective one on a particular news site, it should be the most effective algorithm on other news sites, or at least among the most effectives. We expected that the approaches that performed well on one news website would also perform well on another news site. The algorithms ranged from recommending the most popular news to content-based filtering and collaborative filtering. We participated in the challenge in which a number of recommendation approaches were evaluated. The teams generated recommendations based on the provided information in real-time (within 100ms) and sent them back to Plista, which forwarded them to its partners, i.e. ![]() In 2013, Plista hosted the RecSys News Challenge, in which participating teams received recommendation requests from Plista, including information about the current users and potential recommendation candidates. Plista is an advertisement platform that delivers news recommendations to its partners. conducted the experiments with the news recommender system Plista and the research paper recommender system of Docear (Beel, Langer, Gipp, et al. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |