Open for Innovation
Showing 2 of 2 projects. Click any project card for scope, mentors, and proposal studio.
Mentors: Student: Rytis Kumpa
<p>This project will involve implementing nodes for the KNIME Analytics Platform that would aid evaluating clustering performance, and detecting outliers, in addition to other clustering algorithms.</p> <p>This project will be split into several substages:</p> <p>• Implementation of clustering analysis metrics: Silhouette Coefficient and Davies-Bouldin Index. These will be the first nodes to be implemented and I will need some time to get familiar with KNIME Node development, thus this will take a bit more time. Estimated time necessary: 2-3 weeks.</p> <p>• Implementation of Fast-MCD for outlier detection. Estimated time necessary: 2 weeks.</p> <p>• Implementation of an interactive interface for analyzing clustering performance. Estimated time necessary: 3 weeks.</p> <p>• Implementation of at least one more clustering algorithm: K-means– and/or COD. More will be implemented if spare time is left. Estimated time necessary: 3 weeks.</p>
Mentors: Student: jguertl
<p>The goal of this project is to implement an interactive JavaScript view to enable “Active Learning” within KNIME Analytics Platform. While KNIME Analytics Platform already contains a JavaScript enabled node for the manual annotation of samples (Table Editor), a dedicated active learning node would allow users to interact with a KNIME workflow more intuitively and allow for active learning specific functionality such as a labeling wizard or fast multiobject labeling. After the implementation phase, the person working on the project is encouraged to build active learning examples that involve text or images in order to test implemented node(s).</p>